2009-09-28.log

--- Day changed Mon Sep 28 2009
-!- genehacker [i=genehack@128.62.63.141] has joined #hplusroadmap00:06
fenni'd much rather have a standardized certification test that can be transferred around between various fabratories so you don't have to get re-certified at every location00:10
fenninstead of a 'diploma'00:10
fennthis is re: fab academy00:10
fennalso a computerized format to represent the skills people have could be useful00:11
fenn'who knows how to fix this thing and is currently here?'00:11
NoahjHmm, that's a good idea, I don't think fab academy has a final exam yet, but when it does it should be portable beyond the academy itself00:12
fenni don't mean a final exam, i'd much rather have individual skill-specific exams00:12
fennbecause 'final exam' is so wishy washy anyway and would change from place to place00:13
fennthat's why modern education is useless, you don't know what skills the person actually has00:13
NoahjA fab academy diploma's supposed to represent all the certificates00:16
NoahjAnd eventually there will be fab academies at every fab lab 00:16
fennbut most people don't need or want all of the skills00:16
fennyou'd have to complete everything before getting a diploma00:16
genehackeranyway there are few fablabs00:17
NoahjNo, you can just get a certificate in a specific skill 00:17
fenni just want to know what they're capable of doing, not whether they're a rounded "completely literate person"00:17
genehackerI think they're cool and all but are they successful in their goals?00:17
fennwhat's the point of a diploma?00:18
NoahjA diploma is just all the certificates 00:18
fennbut what is "all"?00:18
NoahjThere's a list of certificates on the fab academy website00:18
fennif you add a new certificat next year do you take away peoples' diplomas until they do that one?00:18
NoahjHmm00:19
genehackerblah blah blah it's just a piece of paper00:19
NoahjThat sounds like an impractical way to go about it00:19
fenn"all the courses some place offered in 2009" is a research project in itself00:19
fennhow will you disseminate the information about what a diploma means?00:20
fennif i have a diploma, how do you know i'm qualified to use XYZ machine no questions asked00:20
NoahjWell, I don't expect a lot to change year-to-year 00:20
fennsee what i'm getting at? i want to just wave my piece of paper and get to work00:21
fennbut the guy responsible for me not getting killed or damaging stuff has to know that i know how to use XYZ machine00:21
fennso he has to look up exactly what this piece of paper means00:21
genehackera lot does change though00:22
genehackernew machines come out, interfaces change, heck maybe new file formats get introduced00:22
NoahjI think we could fix this by printing the list of certificates the diploma covers on the diploma itself00:22
fennok but i still think the whole 'diploma' is just silly00:23
NoahjYeah, re-taking a class every year to certify yourself in the latest kind of laser printer seems a bit like an overkill 00:23
fennalso the point i was originally getting at was that you shouldn't have to take the class to get certified00:23
NoahjI think the diploma aspect is to cater to Boeing and people who want resumes, it's not inherently any better than a collection of certificates 00:24
fennif you can demonstrate you know what you're doing with some standardized test, you shouldn't have to take the class that supposedly prepares you for that test00:24
genehackerI think that is a valid point00:24
genehackerinstead of having certification why not just read the manual00:24
NoahjI think the fab academy will be a bit more hands-on than that, but I really can't say 00:24
fennthere's a lot of stuff that's not in "the manual"00:25
fennalso a lot of procedural knowledge (habit forming)00:25
NoahjA fab academy certification is kind of like a road-test for fab lab tools, as far as I can tell 00:25
fennif you show up in sandals to do welding you fail00:25
genehackeror shorts to a machine shop00:26
genehackeranyway let's forget the fab academy00:27
fennpants just keep you from getting chips in your shoes :)00:27
fenni assumed it was why noah was here00:27
genehackeranyway00:28
genehackerso I'm wondering about something00:28
genehackerthere are these sponges that can make glass from seawater00:29
fenndiatoms?00:29
genehackerno sponges00:29
genehackerhttp://en.wikipedia.org/wiki/Venus%27_Flower_Basket00:29
genehackernow if you can make glass fibers you can can make glass fiber composites00:30
fenninteresting00:32
genehackernow then figure out how to grow glass fibers in a natural polyester matrix and BAM! you're on the way to growing huge living spaceships/monsters00:33
fenni guess00:34
fennmetalloproteins are more interesting00:34
genehackeryeah00:34
genehackerthey are00:34
genehackerquite simple too00:34
genehackerI wonder why there aren't any organisms with metal in them00:35
genehackerIE structural metal components00:35
fenni remember some ant has copper? titanium? protein composite mandibles00:35
genehackerwhoa00:36
genehackerI'd expect copper00:36
fennnow i can't find the article00:37
genehackerI think I found it00:37
genehackerholy shit00:37
genehackeryou're right00:37
genehackerhttp://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6WCB-45N4PCG-P&_user=108429&_rdoc=1&_fmt=&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1026318970&_rerunOrigin=google&_acct=C000059713&_version=1&_urlVersion=0&_userid=108429&md5=b1802b666435c30a6fca71236a557c6800:37
fennmanganese and zinc00:39
genehackerI wonder if it's because there isn't that much available metal in the biosphere00:40
genehackerbut there is now00:40
fennzinc metal isn't bioavailable00:42
fennbesides termites don't eat metal00:43
genehackerthey eat wood00:43
genehackerwhich has tiny amounts of zinc00:43
fennit also has all sorts of other metals00:44
genehackerbut not in large amounts00:44
genehackerhmm...00:45
genehackerthere are metal reducing bacteria00:45
fennapparently pregnant women sometimes get cravings for red dirt and eat it (you can get it in southern restaurants!)00:46
genehackerand reducing metal cations is how you do electroplating00:46
fennfor the iron00:46
genehackerthere is also a disease know as pika00:46
genehackernow that's weird00:46
genehackerpica00:47
fennLOL "My friend has Asperger Syndrome, how can I help?"00:48
genehackerI wonder how that evolved00:48
genehackerand how it happens00:48
fennum, where do you think all that iron comes from?00:49
fennyou have like 1 gram of iron in your body00:49
genehackerno I wonder how it happens in the brain00:54
genehackertime to read up on neural networks00:54
fenni think it's because we know what iron tastes like00:54
fennhm i think google brought back 'dont fuck up my search query with word substitutions' magic plus sign00:57
fenncan you find the sliver you're actually allowed to use? http://www.thingmagic.com/images/frequency_big.jpg01:00
fennthis has the whole thing http://en.wikipedia.org/wiki/Spectrum_allocation01:06
AchiestDragonkatsmeow-afk: / drazak :    a couple of pointers on audio amps ,    use bifet opamps   something like a opa228pa    there is a noticable diference even if spec wise there seems to be no real diference 01:22
AchiestDragonfets / mosfets   like  tubes   are current mode devices  where transistors are voltage mode ,,,  you may see quite a few  audiophile bods saying how good tube amps are copmared to transistor/ic  types 01:24
AchiestDragontheres verry few  true mosfet designs  arround ,, if you do take the time to get all semiconductors on the audio path to be mosfets /fet devices  then you would get one of the few  semiconductor amps that equels the tube amps  on  tone, audio qualaty and give you a nice flat responce in the process 01:27
AchiestDragonalthough a few of the top names in the audio industry  (top qualaty audio wise that is )  did produce mosfet designs  seems that they stoped  the only reason is  the audio mags  just started to proclame that all the semiconductor amps dont beat tubes 01:29
AchiestDragonso even if  mosfets are better  there not tubes so the audiophile's  think there no good because there not tubes ,, genaraly they would not want to show that amp to there mates on  account of the fact its not got nice glowing bottles in it eating up 4* the electric  nomater how good the amps actualy are 01:32
AchiestDragonlike  the same sort of people that would beleve microsocks adviertizing  and think that linux is some pirate version of win95  01:34
-!- genehacker [i=genehack@128.62.63.141] has quit [Read error: 145 (Connection timed out)]01:34
fennmy favorite is the 'plinths' little blocks they use to keep the speaker cord off the ground, to prevent ground loops :)01:37
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has quit []01:37
fenn"it should only contact the ground at one point"01:37
fenni can't figure out who buys any of the crap on crunchwear.com01:41
fenn"ipod enabled" and led illuminated clothes01:41
AchiestDragonhttp://www.whipy.demon.co.uk/10V-PSU.pdf          < alternate way to get the +- v supply for opamps  when you have only a single supply rail 01:43
fennlots of handbags with lcd screens playing various disney movies01:43
AchiestDragonyea   speaker cables there is a diference    on a 100w rms amp   into  8ohm  you are looking at needing a psu that will give about 8A at +- 55v    when playing music its the peek responce that matters   thin speakre cable  would tend to act like resistive current limiting  so not giving the desiered peek responce 01:47
fennwhy 8A?01:48
AchiestDragon4A per chan and over rated enough that when you do get a music peek that causes the caps in the psu to almost discharge  that theres enough left in the psu to recharege the caps while still ensuring the psu voltage does not dip 01:51
AchiestDragonbut for a simple  design that works  with  normal qualaty (ie its often used on commertial  kit    use a  lm356n  single chip 1w audio amp   , ideal for use as a headphone amp  and single supply 01:59
AchiestDragonbut its not studio quality 02:02
fenni am wondering what this has to do with "3d printing" http://www.yankodesign.com/2008/10/03/all-the-cool-kids-wear-air-jordan-prosthetics/02:36
AchiestDragonbecause they use 3d printing to ensure that they get the fit right 02:40
AchiestDragonif you lose a leg  then the stump left would  have a unique shape to the end ,, that prosthetic has to fit to that shape 02:42
fennobviously there are easier ways of doing that02:42
fennlike polymorph for example02:42
AchiestDragonnot practical 02:42
fennapparently making an exact fit is not the right way to do it anyway02:42
fennthe "luke arm" uses this spider shaped gripper that wraps all the way around the stump02:43
AchiestDragonie  you get the  end of the stump scanned and you can just reorder the prosthetic without having to get it rescanned 02:43
AchiestDragonand yea its not an exact fit  and you cant take a mould because there may be  indented surfaces in the stump that would need to be removed in order to get it to slide on to it 02:45
fenni just dont think that's what they did02:45
AchiestDragonso they do a 3d scan  then  modifiy the surface  to get the correct shape  then  plot it to that shape 02:46
AchiestDragonthey do the same when making false teeth ,,, they take a mold ,  the mold gets scanned  they modify the scanned image to get it to fit correctly  then cnc out the plate to that shape 02:47
AchiestDragonknow a guy that makes false teeth ,, hes got a cnc machine that that is accurate to a single micron 02:49
fennoverkill for a tooth02:51
AchiestDragontheres also a guy  not far away that makes  custom  hand grips for sporting rifles ,, again he scans the hand and mods the data to get the grip  usable ,, but fits perfect then even if the user is missing a finger 02:51
fennhow exactly does that work? i mean "scans the hand and mods the data"02:51
AchiestDragonyou take a mould  but  if you just used that then you would need to cut away the outer surface   ie you just want the inner surfaces of the hand that actualy touch the surfce so  you have to edit the scanned 3d model to just get the mating surfaces 02:54
AchiestDragonthen you may need to mod the mating surfaces so that theres no odd bits sticking out 02:55
AchiestDragonits easy to do in cad 02:55
AchiestDragonbut a total pita if you are trying to cut up the original mould to get it the right shape 02:55
AchiestDragonif you make a mistake ediditing it then you got to recall the guy so they can take another mould 02:56
AchiestDragonrather than beeing able to just reload the original scan data and start again 02:57
AchiestDragonoverkill maybe , but  its both an effective way to do it and cost effective  not to metntion  no having a warehouse full of  old  moulds so that you got  a referance incase of a reorder 02:58
fennright02:59
AchiestDragonphisicaly storeing  1000 or so plastic legs  is going to cost quite a bit of  ground rent and heating overheads  on a yearly basis , not to mention the staff to manage the catalogeing of it 03:00
AchiestDragonnot to mention distance problems ,,,  it would be a pain for a person in say  LA having to go to NY to get a mould taken , when there may be a place in LA that can scan the surface and send them the 3d moddel file 03:02
AchiestDragonas to overkill on the accuracy ,,, well better to have the item made to a couple of microns tolarance than gt sued $1,000,000 because by someone who says it dose not fit right  they must of made it wrong 03:07
AchiestDragonand do remeber false teeth are not a one pice mould ,, the teeth part are all individualy cut to shape and mounted on the individualy formed plate 03:11
AchiestDragonso a bit more advanced these days than a guy sat there with a dremmill  individualy hand sculpturing each one from resin  then casting a metal copy of that  03:14
AchiestDragonwhat they do for crowns these days is to take a mould of the tooth before they start ,, big hole in it included ,,  they send that mould off it gets scanned   a guy then edits out the hole on the scan so as to rebiuild the surface ,, computer match it so that the sape fits normal jaw movments and will not bind with other teeth and cnc cut the crown from that   03:18
AchiestDragonthe dentist in the mean time has  got  the tooth  ground down and filled  ins such a way as to take that crown  ready for the 2 to 3 weeks it usualy takes to arrive 03:20
AchiestDragonin engineering  you will find nothing that  comes  close to the  requiremnets that have to use for medical kit   ,  all  over rated and over spec for what you would think is needed ,, and  its needed  for one of  the above mentioned reasons  03:23
AchiestDragoncar  manufacture for example  well  other than making parts that actualy work and fit correctly  theres no manatary tolarances  like the stearing wheel dia can be 10" or 20"  03:25
AchiestDragonthe next major one is   equipment for  food  use ,,  like as for cars  but  materal types  , oil grades and types , have to be good enough , then theres  scientific equipment  that tends to be high tolarance , then medical kit  where its high tolarance and specified material types 03:28
AchiestDragontheres also  space grade kit  like medical but a diferent set of materials and requirements 03:29
AchiestDragongenaral  machining  for  consumer equipment is ususaly only acurate to  about  1mil (imperial)  03:33
AchiestDragoneven if the machines are capable of higher that they have to be anyway   like 0.2 mil would be common 03:34
AchiestDragonyou could start to feel that amount of error if it was a step in the surface of that 03:35
AchiestDragonif  you  can't  ask a blind man  to  03:36
AchiestDragonwith the prosthetics  another aspect that you would probablay overlook is the fact that in contact with it you have living skin ,, that  sweats and can stick to it ,, any error would make that effect uneaven  that can lead to irritation 03:40
AchiestDragonirritation causes swelling that means it nolonger fits , situation gets worse 03:40
AchiestDragonunlike  metal on metal where  that sort of error would naturlay ware down the surfaces  so its not as mutch of a problem 03:43
fenni'm still trying to figure this thing out http://www.yankodesign.com/2008/10/31/embrace-your-femininity/03:47
fennafter several hours i have figured out that stuff on yanko design is not necessarily developed into a real product or even plausible 03:48
fennbut some of it is.. and they never say what's real or not! so frustrating03:49
AchiestDragonwell many companies do put   prof of concept systems there   like  they have been showing  3d laser projection tvs at shows for some time  but still to see 1 of them actualy on sale 03:50
AchiestDragonthey need to  in order to get feedback and hopefully orders so they can assess if they can get funding for production 03:51
AchiestDragonfor example   if you came up with a machine that could open any packet  you put in it  properly ,, then it may be a realy good idea and a realy good system ,, ok so it may cost a couple of milion per machine , so  is it going to be worth making them 03:53
AchiestDragonwell  bet you got a realy limited  customer base ,, i cant think of one place that would need to open enough random types of packets   ie  from a pill bottle to a fraight container  that could justify  having one 03:56
AchiestDragonmaybe with the execption of customs 03:56
-!- ve [n=a@94-193-95-252.zone7.bethere.co.uk] has quit [Read error: 113 (No route to host)]03:57
AchiestDragonat this point the design enginers think   ho good another product idea  one to repackage them after ,, and the accountant heads to the bank and files banckrupsy  03:57
AchiestDragonrambling on i am ,, will stop 04:00
fennok i hope this one isn't a real product http://www.yankodesign.com/2008/09/09/hitch-a-free-ride/04:08
fennhmm two snow crash references in one day.. guess i must read it04:11
-!- ve [n=a@94-193-95-252.zone7.bethere.co.uk] has joined #hplusroadmap04:18
-!- genehacker [i=genehack@wireless-128-62-100-209.public.utexas.edu] has joined #hplusroadmap08:55
-!- genehacker [i=genehack@wireless-128-62-100-209.public.utexas.edu] has quit [Read error: 60 (Operation timed out)]09:06
-!- superkuh [n=hukrepus@unaffiliated/superkuh] has quit [Read error: 104 (Connection reset by peer)]09:53
-!- superkuh [n=hukrepus@24.245.50.49] has joined #hplusroadmap09:54
-!- genehacker [i=genehack@wireless-128-62-174-181.public.utexas.edu] has joined #hplusroadmap09:55
-!- wrldpc2_ [n=benny@ool-ad03fe34.dyn.optonline.net] has joined #hplusroadmap10:00
drazakAchiestDragon: heh, I know exactly how good tube amps, I have a golden tube SE100SE and a fischer x100, both very good tube amps, bipolar mosfets however are very hard to use, they're cranky and need to have the design be made around them10:41
-!- genehacker [i=genehack@wireless-128-62-174-181.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]10:45
drazakAchiestDragon: they also aren't suitable for battery powered amps, too much current draw10:45
AchiestDragoni have built one  some years back ,, i used  some rather large power mosfets ,, was  rated at  400W rms per chan  on 4 chans ,,(about 1.5kw peek music power as the usual " fake blurb spec" that they try to parm you off with )10:46
AchiestDragonits worth doing  on that 10:47
-!- genehacker [i=genehack@128.62.63.141] has joined #hplusroadmap10:51
katsmeow-afki have had bipolar and tubes, i prefer a limited set of bipolars10:51
katsmeow-afkhad an Eico 35w/chan tube, and a lil mono for center channel, and then had a Sansui AU7700 and a Marantz ,, i prefered the Marantz10:52
drazakmarantz makes some nice stuff10:55
drazakI like tubes though10:55
katsmeow-afkthe Eico was a serious room heater, the Sansui had a fine enough preamp section (it's power amp oscillated driving reactive loads), and the Marantz couldn't be beat for power and clarity (tho the 135v that the powerco delivered did beat it into submission -- poof)10:55
katsmeow-afk+/- 80v mains can deliver some kinda brute force10:56
katsmeow-afki spent 2 hrs trying to unload Norton from a computer, finally resorting to manually deleting all references to Norton from the registry10:58
drazakthat's what you have to do, sadly10:58
drazakI remember when norton didn't suck10:58
katsmeow-afkme too, like prior to windoze9810:58
katsmeow-afki still have the Sansui and the Marantz, i stopped all electronics work at the start of the Dog Era, so no controlled psu to plug either into, so i didn't bother replacing their power outputs for the 3nd time each11:01
drazakhehe11:02
katsmeow-afkgot a cheap JVC SuperA i been using11:02
drazakew11:02
drazakyou need to listen to some mcintosh tubes though11:02
katsmeow-afkyeas, no low end, bad components thruout, and no speaker damping11:03
katsmeow-afki knew someone who had a mac11:03
katsmeow-afkhe spend $1000's on each of those things11:03
katsmeow-afki think the only way i'd go back to tubes is if the speakers could be driven directly by the tubes, with no transformer11:04
drazakyeah11:04
drazakI just like tubes11:04
drazak:)11:04
katsmeow-afki could build one again,, i have 14 6AS7G sitting right here, and some 6080's 6SL7's and 2000 assorted other tubes11:06
katsmeow-afknot that i am into tubes ;-)11:06
katsmeow-afkthe sweetest sounding amp i ever heard was a olde vac tube amp with 6L6 outputs (!!!!) and electromagnet speakers11:08
katsmeow-afkit put the oooooo in smooooth11:08
katsmeow-afkya know, you prolly could drive a speaker *enclosure* directly with vac tubes, if you series a mess of 16 or 32 ohm voice coils11:09
katsmeow-afkand enough tubes11:09
katsmeow-afkcapacitiavely coupled them to a class A , with a choke for the plate's dc11:10
katsmeow-afkno transformer11:10
drazakhehe11:10
drazakthe fischer takes 6l6I think11:10
katsmeow-afk6L6 was known to be a dirty amp tho,, or am i thinking of the 6V6 ?11:11
katsmeow-afkiirc, one of the 2 was very nonlinear11:11
drazakoops, EL3411:12
katsmeow-afkhad to use careful design abd lots local feedback11:12
katsmeow-afkwhich is ok with me, i never liked gobs of global feedback11:12
katsmeow-afki'd rather have [clean stage] - [clean stage] - [clean stage]11:13
katsmeow-afksaves on phase margins11:13
drazaktube amp with all smd components but the tubes11:18
katsmeow-afkheh11:19
katsmeow-afksmd 300v caps11:19
katsmeow-afksmd 5w resistors, lol11:19
katsmeow-afki'd hate to see a pcb that had a 5U4 socket mounted to it for long!11:20
katsmeow-afkunless it was fan cooled, of course11:20
katsmeow-afkseen too many of those with glowy plates11:20
katsmeow-afkhave you seen the steam punk watches with vac tube displays, and solid state other guts?11:21
katsmeow-afkthe chips ran on battery power, and you press a button gto turn on hv to the tubes to see the time11:22
drazaknice11:27
-!- splicer [n=patrik@h192n2c1o261.bredband.skanova.com] has joined #hplusroadmap11:37
kanzurehello splicer11:39
kanzurejoshua keena just emailed me 20 SLDPRT files 11:50
kanzurefor the page turning device11:50
splicerhi kanzure11:50
kanzure84 attachments in an email? this guy needs to learn about archives11:51
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has joined #hplusroadmap11:53
genehackerwhoa11:55
genehackerpost that on hey bryan11:57
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has quit [Client Quit]11:57
kanzuregenehacker: uploading now.11:58
kanzurehm. is there a way to make an ink that chemically bonds to a gas? 11:59
kanzureit would be much more effective to just put a book in a chamber, and ramp up the pressure with some gas that would make a magnetic particle bond to the ink on the pages11:59
kanzureand then just do magnetic resonance to scan in the book12:00
kanzureactually i guess it could be liquid12:00
kanzurei'm pretty sure there's magnetic ink somewhere out there..12:00
kanzurehttp://adl.serveftp.org/~bryan/ezreaderfiles.zip12:04
kanzure(6.9 MB)12:04
genehackeryup12:05
genehackerthere is12:05
genehackermagnetic paint at least12:05
genehackeryou could do some weird x-ray stuff12:05
genehackerif you had a very very very fast computer12:05
genehackermost inks already contain iron12:06
kanzureis that so12:06
genehackerbut I'm not so sure about NMRing books to scan them in12:06
kanzureit's faster than flipping pages12:07
genehackeryeah but how do you image a bunch of really thin iron containing layers stacked on top of each other?12:07
kanzurei don't know the layer thickness of MRI12:08
kanzurebut i imagine that it would just be a matter of putting the book on a stage and adjusting it12:08
genehackerit's on the order of millimeters at best12:09
genehackerhmmm...12:09
kanzurehow thick is a page? 500 microns?12:09
genehackerpaper is 0.1 millimeters the best MRI machines are about that12:09
genehackerso it might be doable12:10
genehackerwhat about turning the raw data into to ebook though, that could take some time12:10
kanzurei guess the real trick is doing OCR in less time than it takes to flip a page12:11
kanzureotherwise you're boned12:11
genehackeryeah12:11
genehackerwhat if the book has pictures?12:12
genehackerblack and white pictures12:12
kanzurethere's this algorithm known as image segmentation12:12
kanzurewhere you can separate different portions of a page based off of white space between the regions12:12
kanzuresee here: http://heybryan.org/~bbishop/docs/scripts/segment_image12:12
genehackeranyway mri has 0.1 millimeter resolution at the very very best12:13
genehackerletters and pictures are higher res12:13
kanzureno that wasn't the problem12:13
kanzurethe problem was the thickness of pages12:13
kanzureso to account for that you can put the book on a stage and just have it slowly move up and down until it comes into focus in the MRI machine12:13
kanzurebut hopefully that wouldn't be needed12:13
genehackerwhat has someone tried this before12:14
kanzureprobably?12:15
genehackerhmmm... maybe we could ask the MRI expert in the EE building who's doing hi-res MRI if this is possible12:15
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has joined #hplusroadmap12:33
-!- genehacker [i=genehack@128.62.63.141] has quit [Read error: 145 (Connection timed out)]13:48
-!- genehacker [i=genehack@wireless-128-62-169-215.public.utexas.edu] has joined #hplusroadmap13:58
-!- genehacker [i=genehack@wireless-128-62-169-215.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]15:06
AchiestDragonkatsmeow-afk:  drazak:    have a th331 tube here would make a nice 3kw rms audio amp , but the transformer would be  rather large   not to mention the  other 2 transformers needed for the mains   like the 4.8kv at 1A for the annode and 12v at 60A for the heater 15:14
drazakheh15:15
AchiestDragonthey would be about 50 to 60 kg each 15:16
drazakhaha15:16
drazaknice15:16
drazakI have a 80kg tandberg 15:17
AchiestDragonone thing i learned some time ago  with amature radio equipment  is ,  its posible to home build equipment with the same sort of specs and performance as  high end  broadcast kit for the price of top qualaty amature kit   15:20
drazakyeah, it's crazy15:20
AchiestDragonin audio terms thats  like having say a studio mixing desk as your preamp 15:21
drazakthe most expensive things are like, stepped attenuators and stuff15:21
drazakat 130$ a piece15:21
drazakfor two channels15:21
AchiestDragonbut the commertal kit is made for cost  and featres  rather than a qualaty that realy matters 15:21
drazakthat's one of the things found in hifi stuff that's a bitceh to get yourslef15:21
AchiestDragonactualy    today ,, the most expencive bits are the ADC's and DAC's   used to convert the signal  to digital  then  digitaly mix and process  the data  using  "hardware" in a fpga  15:23
drazakwell, yes15:23
AchiestDragonif you got the ADC sample rate at double the freqancy  then you can directly decode  most radio signals formats in the fpga  using fft and spectrum algorithms  implimented in hardware in the fpga 15:26
drazakheh15:27
drazakI know nothing about DACS and ADC's15:27
AchiestDragonwell the none encripted signals that is 15:27
drazakother than my rio karma has one of the highend ones that they always talk about on audio forums15:27
AchiestDragonwell a adc  converts a  analog voltage   to a digital value that reflects the percentage of that voltage in relation to a referance voltage 15:28
drazakI know what they do :)15:28
drazakI just know none of the details15:29
AchiestDragon:) save me rambling 15:29
AchiestDragonwell   with a referance of 1v    and 8 bits    then 1v on the input gives a binary  255 out , 0.5v on the input gives a binary 128 out 15:30
AchiestDragonthe dac same in reverse 15:30
AchiestDragonto mix a signal    its  input a  plus input b and ouput result to output dac 15:31
AchiestDragonto reduce the volume by half you  divide by 2 15:31
AchiestDragonno analog pots to get dirty contacts and be noisy when moved that way 15:32
AchiestDragonyou dont use pots   linear or rotary optical encoders 15:33
AchiestDragonwith motor overides if you are  wanting to make it like a proper midi compatable mixer 15:35
AchiestDragonbut a pot costs  like  20cents ,, a  optical encoder assembly like that and  your not going to get 20cents back  from $200 per control ,, but there nice 15:36
drazaka nice pot is like 26 bucks :D15:37
AchiestDragonADC   for a  160khz sample rate  48bit  single one is £26 gbp ,,  a 1600 gate fpga module  Â£160  the DAC  about the same   thats about £300  for the digital control   dont know what the current exhange rate is to $ 15:40
AchiestDragoncould eather  control it by a pc or as theres space  configure  softcore micro inside the fpga and have it read a digital front pannel or use some cheap adcs to read pot values 15:42
AchiestDragonreading pot values via a adc is fine   you slow sample them so you lose any noise it may give 15:43
AchiestDragonplus  at  160k samples per second per chan and 48bits  the  signal in and out  of the digital section is better qualaty than anything you could manage in analog  so the rest is down to  what input preamp you may need and whatever final amp and speakers you use 15:45
AchiestDragoncd quality is only 48khz  15:46
AchiestDragonstudio commertial mixers that  use fpgas  these days  as there not produced in enough numbers to justify making an asic , also using fpgas you can flash them an update 15:48
AchiestDragonor reporgam bits of them on the fly 15:49
AchiestDragonsince i have not managed to fund a pcb i need to make ,, been attempting now since jan , im building a cnc machine so i can mill the thing out 15:51
AchiestDragonlike its a pair of  Â£200 full pannels  15:52
AchiestDragononly had funds to cover  Â£100  so far    can get a cnc working for that 15:53
AchiestDragonwell have most of the bits anyway 15:53
AchiestDragongot the z axis  leadscews mounted on it today ,, just the  x and y  leadscrews to fit  then i should be able to  test the drives , and start to allign it all 15:55
AchiestDragonshould have it working by the weekend 15:56
AchiestDragontheres the outer frame and case to make for it also  but got the bits for that also 15:57
AchiestDragonhttp://www.whipy.demon.co.uk/dscf2320.jpg15:58
-!- dira [n=chatzill@de2-as1165.alshamil.net.ae] has quit ["ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]"]16:02
-!- kardan [i=kardan@78.46.51.71] has quit [Remote closed the connection]16:06
AchiestDragonafter some discusion  the machine i was going to build  is to be redesigned from scratch  as its been decided it should have to be able to machine  brass and stainless steel 16:07
kanzurewhich machine?16:07
AchiestDragonand needs to be 5 axis 16:07
kanzurehave you looked at hextatic yet?16:08
AchiestDragonthe one i uploaded to you 16:08
kanzurehttp://fennetic.com/gingery_machines/index.php?hextatic16:08
AchiestDragonho that design 16:08
AchiestDragonyea seen that  but  unless you spend an awfull lot of cash your not going to be able to build one with the strenth needed to do more than plastics 16:09
AchiestDragonie the load on a 20mm dia solid stainless steel bar    the bar would flex at 200mm along more than +- 1mm 16:10
AchiestDragonthats doing aluminium ,,, would need to be a cast block if on stainless steel and some tungsten alloys 16:11
kanzurefenn: hey didn't Part used to inherit from Shape?16:11
kanzureah there's Part.shape16:13
AchiestDragonwell for me yet another project where my design is fine , but the end result is totaly diferent to the plans i drew up  16:16
AchiestDragonwill see if i can find a pic of the old machine its going to replace 16:17
ybitfenn, kanzure: do you have CASROOT and CSF_* definied in your .bashrc?16:20
kanzureyes16:20
kanzureexport CASROOT=~/local/opencascade/OpenCASCADE6.3.0/ros/16:20
kanzureexport CSF_GraphicShr=/usr/lib/libTKOpenGl.so16:20
AchiestDragonhttp://www.whipy.demon.co.uk/dscf2315.jpg16:22
AchiestDragonthe old machine is to heavy and big to get inside ( first floor flat here) so it sits in a little shead  outside ,, where its rusting away and its only realy usable during summer here  16:23
AchiestDragonit also takes a 500W router  so is a bit too noisy to use indoors 16:24
AchiestDragonthe aluminium one  will take a dremel so will be quiet enough and light enough to use here  on the desk 16:25
AchiestDragoncan then concentrate on making a 5 axis machine  that will do metal that can go where the old one is , but will be next summers  project that 16:26
AchiestDragonso the machining time i have avalable is going to be spent making  the custom bits for that 16:26
AchiestDragonso if i get this  machine working  for the end of the month  then  can get the pcbs made intime  to build up  the radio project that i have already acumulated  more than 50% of the bits for 16:29
AchiestDragonlike £800 so far 16:29
AchiestDragonbut still under half what my estimate was 16:30
AchiestDragon:)16:30
AchiestDragonbut still £3000 less than the rhoads and swarts   profesional equivelent  16:32
AchiestDragonwell  i dont work  you can say im retiered  so  got to find something to do to pass the time 16:33
kanzureyou don't need an excuse16:34
AchiestDragonto me its fun ,,  done electronc design for a number of years  profesionaly ,  digital ,  analog and rf and audio  plus  i did engineering at collage many years back also 16:36
AchiestDragonbigest problem you get doing home construction are the tools you genaraly have at hand 16:38
AchiestDragonnow  to do a proper job  you find you need  lathes and a milling machine ,,  that there biggest problem is space ,, a sutable mini workshop is about the size of a 2 car garrage 16:39
AchiestDragonspace i dont have , neather do many others 16:40
AchiestDragonyou cant realy fit a 500kg   1/2 ton  machine on the desk 16:40
AchiestDragonthe router design  you can 16:41
AchiestDragonthe old machine was designed to machine aluminium plate ,, it would do that  but would need a coolant system fitting 16:42
AchiestDragonwith a coolant system fitted  then because of all the crap that you get  you would only be able to use it for metal16:42
AchiestDragonotherwise you block up and contaminate the coolant oil 16:43
AchiestDragonanyway ,, the one i am building atm will manage pcbs  plastics and wood , should be able to do drilling on aluminium sheet 16:46
AchiestDragonits made from  3" by 2"  1/8"thick  aluminium angle ,,, theres a outer frame assembly and covering to make yet , that will brace the machine  tight so it does not flex16:49
AchiestDragonbut its not  far  off strong enough as it is 16:49
AchiestDragonjust a shame i am having to build it as i go allong as that means theres no plans for it 16:50
kanzurethere's enough of us in here to help with the plans16:51
kanzurebut the problem is that we don't have expensive CAD software, only things that support standard formats16:51
kanzure(IGES and STEP are my favorites)16:51
AchiestDragoni will take photos and give approx dimentions 16:51
AchiestDragonas to machining it out  im just clamping  parts together in the right place  then drilling out mounting holes for them , bolt it in  remove clamp and move to next bit 16:53
AchiestDragonjust the way that works well ,, takes  like  2 to 3 days to cad up a part  only to have to go out and hand make it ,   could do eather but no point doing both 16:54
kanzurehow did you determine the flex?16:55
kanzurefor the hextatic design?16:55
AchiestDragonflex ,, any part of the design that may  flex under pressure  and the direction of that flex 16:56
AchiestDragonthe cutting presure can be quite high 16:57
AchiestDragonalso vibration , allong a lenth is going to resanate  so will genearate a oscilating flex 16:58
AchiestDragonexample of how it can affect a machine is think of a washing machine on spin dry , with an imballanced load in it ,, like it tries to bounce arround allong the floor 16:59
AchiestDragonwell the frame structure of that machine would be resinating out of square  because of that 17:00
AchiestDragonlike a slightly bent cube 17:01
AchiestDragonthe hextatic design   well most of that is tube secton frame ,, although there could be a bit of twist in the frame  thats not going to be the problem 17:04
AchiestDragonthe problem with that desing is getting all the angles right on the joins of the frame17:05
AchiestDragonas the acuracy  is limited by the acuracy you get the 3 actuator mountings in relation to each other  and the bed  on all axis's .. 17:07
AchiestDragon6 actuators  not 3 17:08
fennAchiestDragon: an octahedron is not a cube17:08
fenni'm sorry you're just wrong.. build a model out of straws or something17:08
AchiestDragonstraws and putty ,, yea  i know ,, but  get a  aluminium block and manualy drill the holes for fitting the poles into it and i bet you cant manage more than 0.5degree acuracy 17:10
fenni'm not totally certain on how to mount the joint brackets to the frame.. i'm sure there's some way to do it without bending17:10
fenni don't need 0.5 degree accuracy17:10
fennthe whole structure will flex to the right angle17:10
fennanyway we are just talking past each other17:11
AchiestDragonthe distance between the mounting points for the actuators and the head , is verable ok  get one actuator mounted  1mill out of allign  and that actuator path is out by the same amount  so that affects the position and angle of the cutting  head  by the same amount 17:13
AchiestDragonits the lenth determins the relative angle also so if its out on eather  it will affect the machines acuracy 17:14
AchiestDragonyou cant weld it  manualy eather  as it would warp the frame ,   you could cnc weld it if you made a special jig to do it ,,, the way they do make them 17:15
-!- wrldpc2_ [n=benny@ool-ad03fe34.dyn.optonline.net] has quit []17:15
-!- wrldpc2 [n=benny@ool-ad03fe34.dyn.optonline.net] has joined #hplusroadmap17:16
AchiestDragonthe other way is to make a frame then post machine it out the mounting surfaces so they are all in line 17:16
AchiestDragonbut your not going to reproduce the joins needed to make that frame without eather cnc machined joins for the poles or cnc welded , or post machining it  17:19
AchiestDragonwell its not an easy shape to try to weld , hand driling is out , manual adjuster plates for  the mounting points may work  but those actuators are £250 each 17:21
kanzurethis looks like what smari was going for:17:22
kanzurehttp://code.gustavonarea.net/booleano/17:22
kanzure"python" in title and category == "computing" and (rating > 3 or publication_date:year => 2007) and (not software or software:works_on_linux)17:22
AchiestDragonat that price its worth the extra to get a frame made  to do it justice 17:22
AchiestDragonshould of said £250 for the 10" actuators 17:24
fenni don't know what you're talking about17:24
fennacme threaded rod is $10/6ft17:24
ybitkanzure: at leasst your family members use the computer for something somewhat useful, i leave the house with my uncle for 40 mins and i come back to see 'porn-o-rama' on my computer 17:25
kanzureis it good porn?17:26
ybiti dunno, i was too concerned about nasty bacteria and whatever else was left behind on my seat, keyboard, and mouse17:26
kanzurethe obvious solution is to use a proxy to filter all images and replace them with goatse17:27
ybiti think there was a donkey and a panda involved, though ;)17:27
AchiestDragonfenn    threaded rod is the wrong thread pitch and profile to use as a good leadscrew17:27
fennoh i guess i should stop using it as a leadscrew then17:28
AchiestDragonyea cheap and sort of works , but the pitch is too fine and its profile gives too much  friction 17:28
fennbtw i explicitly said "acme"17:29
AchiestDragonsattelite jacks have the right sort of  profile to the threads  but theres no antibacklash 17:29
AchiestDragonthe commertial actuators  they make that are like satelite jacks  like shown on that pic are about £250 like 10X the cost of the satelitte jacks 17:30
AchiestDragona proper machined lead screw  10" long is going to cost about £70 to get machined for you ( there usualy made to order) although they do make standard lenths there still about the same price 17:31
AchiestDragonif you get the leadscrew then you would still need to make that into an actuartor for that design 17:32
AchiestDragonnow i just stipped down 3 sattelite jacks for the leadscrews , but i can mod the nuts to make them antibacklash  but  unfortnatly  you have to cut up the jack to do it , so no way to retrofit antibacklash to the jacks  17:34
AchiestDragonif you could then would be a cheap way of doing that 17:34
AchiestDragonbut as  its not posbile to mod them  it rules them out 17:34
kanzurei wonder if there's a way to traverse the parse tree in booleano17:35
AchiestDragonas to machining  rapidly  your stuck to the maximum  cutting rate of the spindle motor , with a dremil  then you may manage  milling a 1mm cubic aria  per second  from aluminum  17:36
AchiestDragonwhere as with a 1hp motor  you could take  10mm cube per second of a lump of mild steel ,, but that puts the whole machine under the same 1hp worth of stress on each part of the frame while doing it 17:38
AchiestDragonabout the same as trying to bend it by using 2 cars as a vice and drving them together to crush it 17:40
AchiestDragonassuming they both start from where there tuching 17:41
AchiestDragonsomething will give ,, when you get that right  it would be a 20mm dia  milling cutter that should break first 17:42
AchiestDragonother than the actual construction of that frame theres only one draw back of that design 17:44
AchiestDragonspace the ammount of size that frame takes in relation to the amount of workspace it offers  is quite high 17:45
kanzurehttp://piny.be/piny-code/architecture/needed_user_facing_infrastructure/17:46
kanzurepiny is joe rayhawk's ikiwiki+git hoster software17:47
-!- genehacker [n=noko@wireless-128-62-59-229.public.utexas.edu] has joined #hplusroadmap17:47
fenn"CGI frontend to adduser" sounds like a bad idea17:48
fenni wonder if he's actually tried using ikiwiki17:49
kanzureheh17:50
kanzurehe's kind of a masochist sysadmin17:50
AchiestDragonfenn  the http://fennetic.com/gingery_machines/index.php?hextatic type machine  in order for it to be useful for machinining steel  the  supporting bars would need to be  2" dia stainelss steel  poles like on a pillar drill , and it would need blocks macining to join the ends of the tubes  they would have to be  from mild steel , the actuators would need to have the power of mini digger style hydrolic rams  18:20
AchiestDragonshould then have the strenth to take a spindle capable of milling out a 1" wide 1/8" deep  cut from a mild steel block18:22
AchiestDragonso about 2 to 3 hp  spindle motor 18:22
AchiestDragonunfortunatly you would need a 5 axis machine to make the blocks for the frame to start with 18:23
AchiestDragonanything smaller and the macine would bend rather than cut the metal 18:24
AchiestDragonwell bend or self distruct  18:24
fennyou are totally missing the point18:24
AchiestDragonand the point 18:25
fennfirst of all you don't use a 1" endmill18:25
fennsecond, the struts don't bend because all the forces go along the axis of the strut18:25
fennand third you don't need precision 5 axis machined blocks18:26
AchiestDragonyou would do if you wanted to rough cut the surface down  say to make a lip 5mm high all the way arround the edge of a frontpannel ,, like  having to go 5mm deep over  an aria of about 18" by 5"  18:27
genehackerdefine strut18:27
genehackerwhat does object strut refer to in this case18:27
fenna round beam that can change its length18:27
AchiestDragonthe supporting bars of the framework 18:27
AchiestDragonthe bit that adjusts is the actuator 18:27
genehackerthe hydraulic cylinders could be susceptible to buckling18:28
AchiestDragonthere electric linear actuators 18:28
genehackeryeah I know18:28
fennyes buckling is an issue.. that's why the struts are only 18 inches18:29
genehackeranyway triangle top supports should only be in tension18:30
genehackerI am assuming you are making this out of steel?18:30
genehackerperhaps you should use steel with a high carbon content18:30
genehackeroh wait that doesn't affect the modulus18:31
AchiestDragonin tention in one plain  but  its also taking a rotational stress down the axis of the spindle  that will try to twist the whole machine 18:31
genehackerok then dike out the spindle and replace it with a laser beam18:31
genehackerBAM no more rotational torque18:32
AchiestDragonif the frame is too light it would felx ,, the spindle motor would start to vibrate and within a short space of time the whole frame is resonating  18:32
genehackerfortunately you can use the dual cylinders to apply a counter torque18:33
genehackerperhaps one could use active damping18:33
AchiestDragonif the spindle power is 500W then you would be looking at 20mm dia stainless bar at min for those  frame rails 18:33
ybitITK_LIBRARY set as well?18:33
ybitfenn, kanzure18:33
ybit& ITCL_LIBRARY18:34
genehackerfenn do you have cad files of that hexapod?18:34
AchiestDragonyea  laser , plasma cutter , water jet or sand  or even spray gun  ideal  but machining steel with a cutter puts the machine under a lot of stress 18:34
genehackerperhaps you could take advantage of the vibration damping properties of (FILE ERROR) cast iron18:35
AchiestDragonregardless of  type of machine ,,  like taking a hammer and chissle  ,, how much force do you need to give it to chissle your name on the side of a block of steel 18:35
ybitgenehacker: what'd i miss, which hexapod?18:36
genehackerachiestdragon is arguing over this hexapod which I think fenn is working on 18:37
genehackerhttp://fennetic.com/gingery_machines/index.php?hextatic18:37
genehackeroh I know18:37
genehackeruse ECM18:37
genehackerelectrochemical machining18:37
AchiestDragon but my goal is to be able to rapidly mill aluminum and do light milling on steel.         <<<< that is the task is it not 18:37
* ybit hasn't read over the convo, maybe tomorrow since i have a day off then, (yay)18:38
genehackerit's crazy accurate and can cut superalloys18:38
ybitthat's neat18:38
genehackerif you apply MORE POWER you might be able to cut stuff rapidly18:38
AchiestDragonrappid milling   1" at 1/4" deep  per second  , is like  1/4" cube every 1/4 second  18:38
genehackerfenn do you have cad files?18:39
AchiestDragonmore power more stress   , if you think the standard router design  often used is only just capable of machining aluminium with some success  and the requirements for steel are about 10 times the strenth  18:40
AchiestDragonand need a lot higher spindle horsepower to manage it also 18:40
genehackerso how'd we get on this topic anyway?18:40
AchiestDragonthen a 500w router  would requre the frame to be able to take the full load power of that as if jammed  and not flex more than 1 thou 18:41
genehackeranyway why not just make something capable of doing automatic investment casting?18:43
AchiestDragonfor that  on that framework  you would need about 1" dia  bar for the frame ,,, but a 500W router is only going to give you a  1/8" cube per seccond  cut rate at most from steel  18:43
genehackerdid you calculate that out?18:44
genehackerperhaps one could use a material with a higher modulus18:44
AchiestDragonit may take a 20 ton to bend  a 4" dia 1/4" thick  bar , but get a 18" lenth of it  and tap it at the end  it will ring like a bell , and be vibrating more than a couple of mil each way in at the ends in the process 18:45
drazakkatsmeow-afk: that seattle robotics place is a good find, they know shittones about electronics18:45
AchiestDragonhow are you joining the frame poles together fenn 18:46
ybitkanzure, fenn: where is Display.OCCViewer18:58
ybitdid it come in some package that i'm missing because apt-file and locate are clueless18:58
-!- genehacker [n=noko@wireless-128-62-59-229.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]19:16
katsmeow-afkAchiestDragon , if it jammed, i hope your driver electronics are smart enough to shut it down before it got there19:19
katsmeow-afkwhich is why i wanted to use dc, it's easier to control closely19:20
fennframe poles squished at ends and held together with a single bolt at each vertex19:20
katsmeow-afkframe poles = EMT ?19:21
fennsomething like that19:21
katsmeow-afkit's pretty sturdy stuff, and you can beef it some by slipping a smaller size inside the outside one19:22
fenni dont think that would help19:22
fenni did have this idea bout using two pipes with viscoelastic shear fluid between19:22
katsmeow-afki mounted car ac compressors that way :-)19:22
katsmeow-afk30 yrs later, i now know the right way to do it, and can19:23
katsmeow-afki did the squashed emt to mount a golfcart motor in a VW beatle too19:23
katsmeow-afk*that's* some torque19:24
ybitdamn, my pastebin is the first result when googling for Display.OCCViewer 19:29
fennwhy not just use angle iron?19:29
katsmeow-afki didn't have it, nor a way to work with it, 30 yrs ago19:29
fennybit: it should be in python.../site-packages/OCC/Display/19:29
ybitah, found it19:29
ybitfenn: yep19:29
katsmeow-afkand i didn't wann aput more money into something i didn't have a clue would work or not: a VW electric ~197519:29
* ybit was shown to the backroom today. it was neat and depressing at the same time, because i won't be working on anything but one of those items, and it's the only one they gave me permission to even mention to anyone19:37
kanzurehttp://code.gustavonarea.net/booleano/tutorials/evaluable-parsing.html19:38
ybitwe'll be working on a bionic arm here soon, and that makes me happy19:38
kanzurei was going to do something with dependencies today with that19:38
fennyou work at a place with secrets?19:38
kanzurebut now i'm not sure what19:38
ybitfenn: yeah :\19:38
fenndid you sign an NDA?19:38
ybityeah19:38
fennwow19:38
fennwelcome to the dark side ybit19:39
* ybit no likey and yet, i do :P19:39
katsmeow-afkyeas, you can't even complain when your nick isn't on the patent anywhere19:39
ybitlots to learn, sad i can't talk about everything though19:40
kanzurebullshit19:40
ybitit's not super secret kanzure, just stuff like i mentioned above19:40
-!- genehacker [i=genehack@wireless-128-62-169-215.public.utexas.edu] has joined #hplusroadmap19:41
ybitthey're in the process of patenting the arm, that's why i'm not able to discuss that, the other stuff is military tech19:41
drazakok, you guys know how to make mathmatica things and stuff19:45
* ybit doesn't use mathematica19:45
ybitholy fuck, interactiveviewer worked19:45
* katsmeow-afk also doesn't too19:45
drazakor some sort of equation solver19:46
ybitdrazak: go with sage, save the gnus19:46
drazakx=1/(ln(2)*a(b+2c) 19:47
drazaki betcha I could just plug in two of the values and get something nice19:47
ybitdrazak: i'm just messing with you, i only use my trust ti-82 ;)19:47
drazakbut I need to make two of the values so the third is sane19:47
ybittrusty*19:47
katsmeow-afkhell with bug spray, use spray mount adhesive19:48
kanzureybit: what was the fix?19:48
ybitkanzure: that's what i'm trying to figure out19:48
ybitkanzure: i think it was all of CSF_* stuff, in particular, CSF_GraphicShr 19:49
kanzureit would have complained about not having that variable set19:49
ybiti hadn't done that, figured the occ packages would do it19:49
drazakhm19:49
drazakI could use a small value for a and everything else would be set19:50
ybitit did complain about the tcl tck stuff, and it did work after setting that tcl_lib and tk_lib, it could have been any one of those sys env settings19:50
* ybit was just about to specify where OCCViewer.py was, but didn't see the error, so i ran interactiverviewer.py once more to get the error, and it worked19:51
ybitdrazak: i'm not following, you ned to make two of the values so the third is sane?19:52
ybito i c19:53
drazakybit: yeah19:54
drazakybit: I could pick b and c and let a fall where it may for a given value of x19:55
drazakb and c need to be inheriently small to get a large x19:55
drazakand a is inheriently small anyway, cause of what it is19:55
drazaka is a cap, the value for the equation is in farads, b and c are resistors in ohms19:56
drazakwell... I could use a trimpot20:02
drazakbut then I'd have to set the cap so that at near the minimum of the pot it goes to the highest frequency I need20:04
ybitsorry drazak, was fixing the complaints from the lack of CSF_GraphicShr being set, after stating that i had specified it earlier :P 20:05
ybithttp://bayimg.com/image/caemdaacc.jpg20:05
drazakhmmm20:05
fenni just do CSF_GraphicShr='' to get it to shut up20:06
ybiti did export CSF_GraphicShr=/usr/lib/libTKOpenGl.so20:06
ybitdrazak: i'm stumped, try #math? :-\20:07
drazakI'm not stumped20:08
drazak:)20:08
ybitorlly?20:08
drazaknah20:08
fenndrazak: is this a 555 circuit?20:13
drazakfenn: AYE20:13
drazaker, without capslock20:13
fennok how precise/stable does it have to be?20:14
fennif not very you can use a tiny cap = tiny power dissipation20:14
fennlike 1uF20:14
drazakI'm using it as a squarewave function generator for testing headphone amps and testing an oscilloscope20:15
drazakI was actually considering something like 100nf20:15
fennok close enough20:15
drazakneeds to be fairly stable20:15
drazakbut not perfect20:15
fennbtw most capacitors are like +-10% so if your frequency is important, best to use a trimpot20:22
drazakyeah20:22
drazakI know20:22
fennbut if your scope is not calibrated i don't really know how to calibrate the 55520:22
kanzurei wonder if booleano can deal with recursive expressions. i.e., parse_manager.parse("book:title is 'Introduction to Physics' and book:author is " + some_other_parse_tree)20:22
fennunless you have an LCR meter20:22
drazakI could buy a 1% cap20:22
drazakthen it'd be fairly close20:23
fennhow about outputting something from a computer20:23
drazakcomputers are good a sinewaves20:23
drazaknot good at flat waves20:23
fenntoo much filtering on audio out?20:23
drazakyeah20:23
drazakit looks like a garbled mess out of all but the best soundcards20:24
fennanyway it will get you the right frequency so that you can calibrate other stuff20:24
drazakmaybe20:24
drazak:D20:24
drazakI could get it in range with a 1% cap20:24
* ybit needs a smaller cad file, the hemi engine step file won't even display, grr.20:29
ybitand just when i say that, it displays20:29
drazakfenn: trimpots go 0-value right?20:34
fennsome of them, but usually it's part of a switch and not a smooth transition20:34
drazakthat's ok20:34
drazakwhat's the usual minimum value?20:34
fennusually it's open circuit though20:34
fennybit: http://adl.serveftp.org/lab/fenn/lego_stepmodels.tgz20:35
ybithttp://ybit.ath.cx/images/iv2.jpg :: that's me trying to rotate the hemi20:38
ybitty you fenn 20:40
-!- genehacker [i=genehack@wireless-128-62-169-215.public.utexas.edu] has quit [Read error: 110 (Connection timed out)]20:41
fennyw welcome ybit20:41
ybit:P20:42
ybitit works nicely with simple models, what a relief 20:42
ybithttp://ybit.ath.cx/images/iv3.jpg20:44
ybitthe cpu usage spike is from scrot grabbing a screenshot20:45
kanzurehttp://ybit.ath.cx/images/which_street_nashville.jpg20:47
kanzurei think i'll go straight20:47
ybithehe20:48
ybitthat dir is about to increase significantly with sci-fi/futuristic drawings, downloaded quite a bit of those type of images yesterday when brainstorming over the layout on the propery 20:51
fennohboy20:51
fennyou know maybe you should sort your images a bit20:52
flamootall eyes are on the aura of electronic intelligence that shines around the transcendental object at the end of time or, "Singularity"20:56
kanzurewhat the fuck20:56
flamootall eyes on deck20:57
flamootlet's look at it together20:57
* flamoot gazes at the nascent AI's20:57
fennom mani padme hummmm20:58
kanzurehttp://adl.serveftp.org/~bryan/pie.py i wrote this thinking that it would be only a quick next step to do a proper Pie.build method but i was wrong :(21:41
ybitanyone used http://www.hpfem.jku.at/netgen/ ?21:43
kanzureyes21:43
ybitsupposedly good for grid generation21:43
kanzuretetgen too21:43
kanzuregmsh also doesn't suck21:43
ybitso what visualization tool was decided on for skdb?21:49
ybitwas it this: http://www.shocksolution.com/microfluidics-and-biotechnology/python-vtk-paraview/21:50
kanzureno21:53
kanzurerun paths.py21:53
kanzureybit: you have time for a walkthrough?21:55
ybityeah sure21:55
kanzurehm i seem to be having some local problems21:55
kanzure../../../libdrm/intel/intel_bufmgr_gem.c:721: Error preparing buffer map -1349728242 ((null)): Invalid argument .21:55
kanzurepython: vbo/vbo_save_api.c:216: map_vertex_store: Assertion `vertex_store->buffer' failed.21:55
kanzure(i've been having some issues with my graphics card)21:56
kanzurewell anyway, just run paths.py in skdb21:56
ybitright, i have a few problems of my own to fix21:57
ybitenv variables again21:57
kanzurefenn: shape_volume has been deleted apparently?21:58
kanzure(running some of the unit tests)21:58
ybitODD22:00
ybit    import skdb22:00
ybitImportError: No module named skdb22:00
ybit*** glibc detected *** python: double free or corruption (!prev): 0x09d15ba8 ***22:00
ybit^C22:00
ybittogetic:/home/heath/projects/skdb# echo $PYTHONPATH22:00
ybit:/usr/local/lib/python2.6/dist-packages/OCC:/usr/lib/python2.4/site-packages/wx-2.8-gtk2-unicode/wx/:/usr/local/lib/python2.6/dist-packages/OCC:/usr/lib/python2.4/site-packages/wx-2.8-gtk2-unicode/wx/:/home/heath/projects/skdb/22:00
kanzurethat last item in your path should be one directory up22:01
-!- genehacker [i=genehack@wireless-128-62-38-144.public.utexas.edu] has joined #hplusroadmap22:02
ybitah22:02
ybithehe, like the output from the cli22:06
ybitwhy are we using something that's not even in debian (python-graph)?22:06
ybitso what was that, it looked like a tube with a flag hanging it from it :P22:06
ybitoh, an arrow, and it quit doing whatever it was doing because of an error22:08
ybitmaybe i should pastebin some of the relevant cli history? 22:08
ybit  File "/home/heath/projects/skdb/core/interface.py", line 140, in connect22:20
ybit    assert self.interface1.part in cgraph.dict().values()22:20
ybitAssertionError22:20
ybithmm22:21
ybitwhat's the error is my Q22:21
ybiti need more thatn just assertionerror22:21
ybit/usr/lib/libcgraph.so.422:21
kanzureyou probably don't have cgraph22:24
kanzurealso, it turns out pythonocc doesn't compile with swig-1.3.40 and only likes swig-1.3.3622:25
kanzurethat's a funky error anyway, what is cgraph doing in core/interface.py ?22:26
fennsome open manufacturing project context stuffs: http://blog.p2pfoundation.net/crisis-at-the-factor-e-farm/2009/08/1022:27
fenni do wish there were other development sites besides Factor E Farm22:28
fennby "site" i mean physical location where stuff is happening22:28
fennare there others and I just don't know about it? N55 is probably one of the few others and that kinda puttered out after one of the people died22:28
ybitthe beginning of a log while getting paths.py working22:30
ybithttp://pastebin.com/f72635dbc22:30
ybitafk, forming a corporation on 10/28, 10:28 :P22:32
genehackerybit what arm? robot arm?22:33
ybitgenehacker: prosthetics22:33
ybitoh damn, it's 9/28, well that ruins everything22:33
genehackerall I care about is if it hooks up to the nervous system23:29
genehackergood luck23:29
genehackerhope it works23:29
ybitthanks, it should be a somewhat fun learning exp.23:43
kanzureSep 28 23:34:51 pikachu kernel: [ 5518.842406] python[10186]: segfault at 30 ip afed775d sp bf8b8f70 error 4 in i965_dri.so[afe9e000+268000]23:43
ybitgrr, kanzure, what was the name of the book you and i both mentioned yesterday? it was named ~"Nano__ & Micromanufacturing"23:47
ybitcan't find it in the logs23:47

Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!