#emc | Logs for 2008-03-10

[00:02:40] <fenn> one could do some interesting things with a wacom tablet and hal_input. it's a 5-axis input device
[00:06:14] <JymmmEMC> fenn: how do you get 5 axis from a tablet?
[00:08:12] <Duc05> the eraser ;)
[00:08:16] <fenn> it senses xyz and two tilts
[00:08:35] <fenn> the 'pressure' is just Z within some threshold.. the nib flexes and that changes the Z position
[00:08:57] <fenn> you can wave it around in the air above the pad and get position data
[00:10:47] <JymmmEMC> it's the waving kludgy at best?
[00:11:01] <fenn> no, it's high resolution. good stuff
[00:11:46] <fenn> i'm playing with wacdump which shows the raw output
[00:20:40] <JymmmEMC> cool
[00:29:32] <jmkasunich> cradek: diesel injection pumps are one of those things where "fixing" it might just make it worse.
[00:30:10] <jmkasunich> only you know the situation with your generator (what it cost, what its worth to you, etc.), but if I determined that I had an injection pump problem, I'd probably take it to a pro
[00:30:36] <jmkasunich> if I had picked it out of a dumpster and had nothing to lose, I might mess around with it myself
[00:30:41] <owhite> hey folks. I'm trying to run a dwell with a M64 command. Doesnt seem to work. "M64 P2 G4 P.2" gives an error.
[00:31:02] <cradek> like you discovered, you can't have two Ps on one line
[00:31:12] <owhite> indeed I did. any suggestions?
[00:31:15] <jmkasunich> g-code doesn'
[00:31:17] <jmkasunich> oops
[00:31:20] <cradek> only don't do that
[00:31:29] <cradek> use two separate lines
[00:31:39] <owhite> oh, sorry.
[00:31:40] <jmkasunich> g-code doesn't look at the order of things on a line, so you simpley CAN'T have two if ANY letter on a line
[00:31:52] <owhite> dint know it was that simple.
[00:32:21] <jmkasunich> I wonder how hard it would be to spit out an error for any duplicate letter found?
[00:32:35] <cradek> I'm sure it does that
[00:32:49] <owhite> its does. i got....
[00:33:07] <owhite> "multiple p words on one line"
[00:33:16] <cradek> there you go :-)
[00:33:20] <jmkasunich> oh
[00:33:38] <cradek> jmkasunich: yeah I'm about done trying
[00:33:42] <jmkasunich> I assumes since you were asking here that you got a meaningless message, or none at all
[00:33:51] <cradek> frustrating though.
[00:34:08] <owhite> but if I put the G4 on the next line, it will perform a dwell on the previous command?
[00:34:26] <jmkasunich> I dunno what M64 is
[00:34:29] <cradek> m64 turns on an output
[00:34:34] <cradek> you can delay before it, or after it
[00:34:50] <owhite> great.
[00:35:01] <jmkasunich> it actually is much less confusing if you put the G4 either before or after, if you put it on the same line, which do you want?
[00:35:23] <lewin1> lewin1 is now known as lewing
[00:36:41] <owhite> I want to turn on my laser, and it dwell. So I'm guessing I'll do M64 P2\n G4 P.2
[00:36:51] <jmkasunich> yep
[00:40:42] <Unit41> i kick ass
[00:40:46] <Unit41> my lathe is allmost done
[00:40:54] <Unit41> 1 more hour
[00:41:14] <Unit41> made it out of a fencing pannel
[00:41:17] <Unit41> and canadian tire parts
[00:59:38] <Ito-Brazil> Hello EMC masters, please i need some help
[00:59:59] <jmkasunich> again?
[01:00:02] <Ito-Brazil> What is "joint 2 following error"
[01:00:08] <Ito-Brazil> yeah lol !!!!
[01:00:41] <jmkasunich> that means EMC asked a motor (or a stepgen) to go faster than it can actually go
[01:00:52] <Ito-Brazil> I am about to use 2.2, but this error ocurr when i try to run the example
[01:01:10] <jmkasunich> for steppers, you need to make sure that STEPGEN_MAX_VEL and STEPGEN_MAX_ACCEL are about 10% higher than MAX_VEL and MAX_ACCEL
[01:01:24] <jmkasunich> (I may have those names mis-spelled, but its something like that)
[01:02:02] <Ito-Brazil> i understood thanks for Ur time, i will check about!
[01:02:20] <Ito-Brazil> I can (or... may I ?) come back again...
[01:02:38] <Ito-Brazil> Really tks!
[01:05:59] <fenn> i wonder why hal_input doesn't see any rotation data: http://rafb.net/p/r4zIRa11.html
[01:13:19] <Ito-Brazil> Hello there: i am using the stepper_mm.ini on stepper directory
[01:13:40] <Ito-Brazil> there is no stepgen_maxvel, only stepgen_maxaccel
[01:14:27] <jmkasunich> oops, right - my mistake - only stepgen_maxaccel is required
[01:14:45] <jmkasunich> it needs to be slightly (10% or so) higher than max_accel
[01:15:07] <Ito-Brazil> oke tks its 520 and accel is 500, i will change to 560 and try agin
[01:15:18] <jmkasunich> wait a sec
[01:15:25] <Ito-Brazil> im heare
[01:15:33] <jmkasunich> you are using one of the sample configs? and not your own config?
[01:16:04] <Ito-Brazil> I made a copy and changed something
[01:16:14] <jmkasunich> did you change base period?
[01:16:25] <Ito-Brazil> cos i was using the 2.05 version
[01:16:29] <Ito-Brazil> nops
[01:16:35] <jmkasunich> ok
[01:16:47] <jmkasunich> the sample configs should work, if you don't change anything
[01:16:47] <Ito-Brazil> base period... i will check on old .ini
[01:17:08] <jmkasunich> you might want to try the unchanged sample config, if it works, then look at your changes
[01:17:15] <Ito-Brazil> it worked, but im using the phase-drive config
[01:17:26] <Ito-Brazil> oke i will try
[01:19:35] <Ito-Brazil> the unchanged worked
[01:20:38] <jmkasunich> things that can slow down stepgen and cause following error:
[01:20:46] <jmkasunich> 1) raising base period
[01:21:01] <jmkasunich> 2) raising step_len or step_space
[01:21:12] <jmkasunich> 3) lowering stepgen_max_vel or stepgen_max_accel
[01:21:31] <jmkasunich> 4) I'm not sure if there are any more
[01:22:11] <Ito-Brazil> i see, thanks a lot dude! Im testing base period to 150000
[01:22:32] <JymmmEMC> yeow!
[01:22:51] <Ito-Brazil> lol too long ?
[01:22:59] <jmkasunich> yes, probably
[01:23:12] <Ito-Brazil> same errora
[01:23:16] <jmkasunich> that list of things is things that are BAD, that might cause the error you are seeing
[01:23:19] <Ito-Brazil> back to original
[01:23:51] <jmkasunich> if unchanged worked, and changed doesn't work, look at the changes
[01:28:10] <Ito-Brazil> oke, but look: the original was DIR-STEP config, I use Phase-Drive drives. I just changed core_stepper.hal and standard_pinout.hal
[01:29:00] <jmkasunich> well, you need to look at those changes
[01:29:21] <jmkasunich> or do a diff and pastebin the diff, if you want someone else to look at them
[01:30:03] <Ito-Brazil> nice, i will pastebinn those files, thanks a lot!
[01:30:24] <jmkasunich> pastebin the diff, please
[01:32:14] <Ito-Brazil> hm ONLY the diff ?
[01:32:36] <Ito-Brazil> http://pastebin.com/d74d9921a
[01:32:44] <Ito-Brazil> thats my stepper_mm.ini
[01:33:46] <jmkasunich> I don't want to dig through your entire files
[01:34:05] <jmkasunich> if one config works, and another doesn't, then the problem must be due to the difference between those two configs
[01:34:05] <Ito-Brazil> oke
[01:34:21] <jmkasunich> I'm willing to look at that difference and try to see what is the problem
[01:36:37] <Ito-Brazil> http://pastebin.com/m2e281d1c
[01:36:47] <Ito-Brazil> my changes at core_stepper.hal
[01:37:15] <jmkasunich> why did you disconnect the enable pins?
[01:37:32] <jmkasunich> (lines 12-16 in the pastebin)
[01:38:57] <jmkasunich> try uncommenting those lines
[01:39:23] <Ito-Brazil> but... i gess it will give annother error
[01:39:29] <Ito-Brazil> cos this pin is used
[01:39:49] <Ito-Brazil> phase-drives to 3 axis uses 12 pins, all LPT out pins
[01:40:52] <jmkasunich> lines 12-16 refer only to internal HAL pins, they do not go to the parport
[01:41:11] <jmkasunich> those signals go from EMC's motion controller to the software step generator, to turn the stepgen on
[01:41:17] <jmkasunich> without them, you never get any steps
[01:42:47] <Ito-Brazil> http://pastebin.com/m58aad698
[01:42:57] <Ito-Brazil> thats my standard pinout changes
[01:43:11] <jmkasunich> that looks fine
[01:43:32] <jmkasunich> I'm about 99.99% sure your problem is lines 12-16, the enable signals
[01:43:41] <Ito-Brazil> cool!
[01:43:50] <jmkasunich> remove the # on those lines
[01:43:50] <Ito-Brazil> but, not success
[01:44:07] <jmkasunich> same result, or a different failure?
[01:44:29] <Ito-Brazil> same error
[01:45:00] <jmkasunich> do you get any movement at all, or does the following error happen the instant you try to jog?
[01:45:12] <jmkasunich> does it happen on all axis, or only one?
[01:45:32] <Ito-Brazil> I will make a hard test right now
[01:47:10] <Ito-Brazil> Same error on JOG, a pressed the button and BOOOM! error
[01:47:43] <fenn> why were the enable signals commented out?
[01:47:51] <jmkasunich> fenn: misunderstanding I think
[01:48:12] <jmkasunich> Ito-Brazil: while EMC is running, open another shell, and run this command:
[01:48:15] <jmkasunich> halcmd show sig
[01:48:21] <jmkasunich> then pastebin the result of that
[01:54:43] <Ito-Brazil> http://pastebin.com/d4677caee
[01:57:09] <jmkasunich> I thought you uncommented the Xen, Yen, Zen lines?
[01:57:26] <jmkasunich> there are no Xen, Yen, Zen signals in your "halcmd show sig"
[01:57:27] <Ito-Brazil> hm i gess
[01:57:58] <jmkasunich> so either you didn't uncomment, or you didn't save after the edit, or something like that
[01:58:19] <Ito-Brazil> it was not enabling the EMC to run
[01:58:39] <Ito-Brazil> I comented those and EMC opened... I wil try uncomment
[01:59:10] <jmkasunich> a while ago, I asked you to uncomment those lines, and you said you did, but it didn't work. Then I asked you if it had the same error as before (following error) and you said yes
[01:59:24] <jmkasunich> now you are saying something completely different
[02:02:58] <Ito-Brazil> i wil check all files again about Xen. Yen and Zen too?
[02:03:43] <jmkasunich> you must enable the step generators
[02:04:48] <jmkasunich> if uncommenting those lines causes emc to stop with an error message, then tell us what the error message is, and we will fix that problem
[02:05:17] <fenn> * fenn guesses multiple writers to a pin
[02:05:25] <Ito-Brazil> EMC runs now, but same error ocurs
[02:05:31] <Ito-Brazil> i will pastebin something
[02:05:52] <Ito-Brazil> joint 2 following error
[02:05:55] <Ito-Brazil> same eror
[02:05:57] <Ito-Brazil> wait a sec
[02:06:51] <jmkasunich> pastebin the output of "halcmd show pin stepgen"
[02:07:25] <jmkasunich> that will show you what (if anything) is connected to stepgen.0.enable
[02:08:00] <Ito-Brazil> # net Xen => parport.0.pin-01-out
[02:08:05] <Ito-Brazil> its commented
[02:08:23] <jmkasunich> good - you don't need enable to go outside the PC throught the parport
[02:08:31] <jmkasunich> but you do need enable to go to the stepgen
[02:08:40] <Ito-Brazil> oke
[02:12:14] <Ito-Brazil> http://pastebin.com/md22b0
[02:12:48] <Ito-Brazil> and a plus error:
[02:12:52] <Ito-Brazil> need to be enabled, in coord mode for linear move
[02:14:19] <fenn> you lost some characters: http://pastebin.com/md22b088
[02:14:19] <jmkasunich> that pastebin is empty
[02:14:48] <Ito-Brazil> http://pastebin.com/md22b088 are U sure?
[02:15:14] <jmkasunich> the first time you pasted, we only saw md22b0, the 88 from the end was missing
[02:15:36] <jmkasunich> ok, look at line 7 of the pastebin
[02:15:49] <jmkasunich> " 5 bit IN FALSE stepgen.0.enable"
[02:15:57] <jmkasunich> as long as that is false, your motors will NEVER move
[02:16:05] <Ito-Brazil> hehe
[02:16:19] <jmkasunich> you MUST connect some signal to that pin in HAL
[02:17:07] <jmkasunich> "net Xen axis.0.amp-enable-out => stepgen.0.enable" is what you want to do
[02:18:08] <Ito-Brazil> sorry, i didnt understood
[02:18:28] <Ito-Brazil> i need to change this line on standard_pinout.hal?
[02:18:41] <jmkasunich> lines 12, 14, and 16 of http://pastebin.com/m2e281d1c need to be uncommented
[02:19:50] <jmkasunich> then line 7 of http://pastebin.com/md22b088 will look like "5 bit IN TRUE stepgen.0.enable <== Xen"
[02:20:35] <Ito-Brazil> <== ?
[02:20:43] <Ito-Brazil> not only <= ?
[02:21:48] <Ito-Brazil> i have much thunders and hard raining now... I gess I will turn off the computer etc..U know 3rd world has not muc protections...
[02:22:05] <fenn> good night
[02:22:11] <Ito-Brazil> Thanks a lot, it was a great light on the tunnel
[02:23:21] <fenn> i wonder how a typical bug session will go when we have a graphical hal configurator
[02:24:23] <owhite> anyone know why I'm getting "M code is greater than 199" when the call I'm making is to M102?
[02:25:32] <SWPLinux> no
[02:25:47] <fenn> what does the surrounding code look like?
[02:26:04] <owhite> its one line...
[02:26:07] <owhite> M102 95
[02:26:15] <SWPLinux> heh
[02:26:20] <SWPLinux> that's M10295 to EMC
[02:26:21] <jepler> because thst's the same as: m10295
[02:26:24] <jepler> which is bigger than 199
[02:26:28] <SWPLinux> you need to use M102 [95]
[02:26:34] <owhite> ha!
[02:26:36] <jepler> no, you need M102 P95
[02:26:38] <cradek> M* take P and Q words
[02:26:41] <SWPLinux> with or without spaces :)
[02:26:50] <SWPLinux> oh, that's true too
[02:26:55] <owhite> jepler. by the way this is the implementation of the code you gave me.
[02:27:01] <owhite> I'll try P95
[02:27:02] <jepler> which code?
[02:27:26] <owhite> oh I askd for something a while back , it takes a number, converts it to bits.
[02:27:39] <fenn> thanks for whoever fixed mdi to select the whole line on up-arrow
[02:27:41] <owhite> you're just so prolific you cant remember all the great code you make :-)
[02:29:13] <fenn> owhite: is this a hal comp?
[02:29:26] <fenn> or a python script?
[02:29:44] <owhite> bash script.
[02:29:57] <owhite> and its working.
[02:30:06] <fenn> gratz
[02:30:36] <tomp2> in that cnc turret posted , what is a .sat file? http://www.cnc-projects.de/CNC-Revolver.html
[02:31:25] <fenn> ACIS according to wikipoogle
[02:32:19] <SWPLinux> http://filext.com/file-extension/sat
[02:32:53] <jepler> http://local.wasp.uwa.edu.au/%7Epbourke/dataformats/sat/sat.pdf
[02:33:20] <tomp2> thx, some French std for ascii description of 3D model, cool
[02:38:25] <ALS> cradek: what JMK said about injection pumps they realy are,nt that complex, the only thing is if you need to calibrate them you need expencive test stands but if its only a delivery valve that is like changing a spark plug
[02:39:18] <tomp2> re: the ACIS format, outside of the word 'portmanteau' i see only english
[02:39:43] <fenn> portmanteau is an english word too
[02:40:10] <fenn> er, only an english weard: would be portmanteaux
[03:07:29] <ALS> cradek: pump pic yanmar http://i272.photobucket.com/albums/jj175/ALS_60/Image3.jpg
[03:08:12] <ALS> http://i272.photobucket.com/albums/jj175/ALS_60/Image2.jpg
[03:09:04] <ALS> not sure if this is like yours or not
[03:09:24] <ALS> good night
[16:03:00] <BigJohnT> I just invented the vaccum wrench
[16:03:27] <archivist> er what!
[16:05:02] <BigJohnT> a vaccum wrench
[16:05:07] <BigJohnT> it sucks
[16:05:10] <BigJohnT> no kidding
[17:14:25] <fenn> a PLL comp would be useful for other things too
[17:15:24] <fenn> i suppose one could also "fake" the index pulse with a counter component in HAL
[17:21:51] <SkinnYPuppY> I think I've googled upon the dumbest linux thing ever.
[17:21:55] <SkinnYPuppY> http://www.strangehorizons.com/2004/20040405/badger.shtml
[17:25:34] <fenn> yeah but can it dance?
[17:25:59] <SkinnYPuppY> Unfortunately yes, it was off this link . http://www.badgerbadgerbadger.com/
[19:21:19] <cradek> heh, I like jmk's dog pics
[19:21:24] <cradek> sproing!
[19:27:31] <skunkworks> happy dog
[20:28:02] <maddash> arghhh1!!!!!!
[20:46:52] <BigJohnT> sounds bad maddash
[20:57:52] <gene> hi guys, is Steve S. around?
[21:00:03] <gene> I just checked the wiki, and there is no mention of the xenable signal that the search can find & I need some advice.
[21:26:32] <cradek> gene: he shows up very rarely. If you need him maybe you will have to call or email
[21:30:56] <kanzure> Is the owner of frennetic.fennetic.net in here?
[21:30:59] <gene> Humm, I'm standing here trying to figure out why I'm not getting any signals on parport pin 14 and 16, the two signal that supposedly run a spindle. They seem to be fixed at about 45 millivolts.
[21:31:04] <kanzure> fenn: You alive?
[21:31:46] <gene> xylotex interface & cable. .hal looks sensible.
[21:32:18] <kanzure> fenn: What is this 'hexapod architecture' ?
[21:33:18] <skunkworks> kanzure: http://en.wikipedia.org/wiki/Stewart_platform
[21:34:18] <gene> Could this be a wrong parport config in the bios problem?
[21:34:41] <kanzure> I don't understand, fenn described it as 'the hexapod architecture is to milling machines as the Bucky-dome was to building architecture (in civil engineering)'
[21:34:49] <kanzure> and I don't see indications of that on the Stewart platform page there
[21:36:29] <skunkworks> That would be a fennism.. He is comparing the hexapod(stewart platform) to (because of its use of triangles I assume) bucky-dome.
[21:36:56] <kanzure> ah, just architecture?
[21:36:58] <kanzure> not in functionality?
[21:37:38] <skunkworks> most likely because it is 'just a better way of doing it' :)
[21:38:15] <kanzure> http://fenn.freeshell.org/retrofest/default.html
[21:38:16] <kanzure> Ah, here we go.
[21:38:19] <kanzure> Hard bastard to find.
[21:38:45] <kanzure> Heh, his bookmarks
[21:39:03] <kanzure> yeah, definitely an Opera user
[21:42:04] <kanzure> heh, a surfraw user
[21:42:22] <kanzure> btw, if anybody wants a copy of the CNC/gingery wiki that once existed, I have a mirror now :)
[21:59:32] <BigJohnT> kanzure: what is it?
[21:59:43] <kanzure> BigJohnT: A wiki?
[21:59:47] <fenn> opera? bah. dillo all the way!
[21:59:51] <BigJohnT> the mirror
[22:00:24] <fenn> the gingery wiki is now here: http://fennetic.net/machines/
[22:00:51] <fenn> however i may split the gingery part into its own thing, if sufficient motivation arises
[22:02:04] <fenn> hexapod is like a geodesic dome because it has efficient use of materials through triangulation and even force distribution
[22:02:49] <fenn> no bending in structure so elasticity is proportional to length of beams, instead of length cubed
[22:05:35] <fenn> also it is made of multiple identical parts for easy transportation and manufacture
[22:06:08] <fenn> we came up with a new idea yesterday that is a tensegrity hexapod
[22:12:45] <fenn> damn, guess i can't make a fractal in inkscape using clones :)
[22:19:05] <kanzure> fenn: Hi there.
[22:20:15] <kanzure> fenn: Is there an export button on that page?
[22:21:54] <fenn> http://fennetic.net/sketches/hexegrity.png
[22:22:58] <fenn> hexapod tensegrity structure. if you add springs to the bottom three cables you can counteract gravity (and if you add more force than gravity it will add preload tension to the cables which is very useful)
[22:24:05] <kanzure> I don't see the functional usefulness of that.
[22:24:44] <fenn> there are lots of uses
[22:25:01] <fenn> for example you could squirt out concrete and shape it with paddles
[22:25:24] <fenn> or for plasma cutter/waterjet
[22:26:07] <fenn> basically anywhere a traditional hexapod is too heavy, expensive, hard to transport
[22:26:32] <fenn> if you dont need super duper high rigidity and precision
[22:26:51] <kanzure> fenn, have you ever come across http://reprap.org/ ?
[22:27:11] <fenn> yep
[22:27:32] <kanzure> they are near total self-replication
[22:27:38] <fenn> yeah, sure..
[22:27:41] <kanzure> I have found a way to do the transistor replication via semiconductor nanoparticle synthesis
[22:27:57] <kanzure> the main problem in the mean time is metal fabbing
[22:28:10] <fenn> if you consider self-replication assembling a bunch of off the shelf parts, well, where does that get you
[22:28:10] <kanzure> it looks like you have some ideas on your wiki for this?
[22:28:17] <kanzure> nah, it's not off-the-shelf
[22:28:28] <kanzure> most of the components can be made by its own tools
[22:28:46] <kanzure> except the raw materials, but that's acceptable
[22:28:52] <fenn> they use precision ground steel rods and stepper motors and electronics and a plastic squirter nozzle and threaded rod and.. and..
[22:29:07] <kanzure> grond steel rods? That's something that a CNC should be able to do
[22:29:13] <kanzure> stepper motors are just wires + insultors + a magnet, right?
[22:29:16] <kanzure> and another rod
[22:29:24] <fenn> why doesnt the reprap "just make" them then?
[22:29:43] <kanzure> the plastic squirter nozzle I have no clue about, I suspect that since it's made out of plastic they should be able to make it with their plastic-mold-arm ?
[22:29:45] <fenn> because all it can do is squirt out plastic
[22:29:52] <fenn> and squirted plastic is good for shit
[22:30:07] <kanzure> adding another arm isn't hard
[22:30:10] <kanzure> and the extra arm can be made out of plastic
[22:30:30] <kanzure> and this extra arm can, say, print out transistors (on the mm scale, not micrometer and definitely not nanometer)
[22:30:33] <fenn> you can't make a thermoplastic squirter nozzle out of thermoplastic, that should be obvious
[22:30:44] <kanzure> what's the squirter nozzle made out of ?
[22:30:46] <kanzure> metal?
[22:31:07] <fenn> * fenn googles
[22:31:09] <kanzure> heh :)
[22:31:14] <kanzure> ./join #reprap
[22:31:30] <fenn> looks like brass wrapped in teflon
[22:31:54] <kanzure> hm, teflon is difficult
[22:32:37] <fenn> still, at least reprap is designed for economy and with the idea of home-fabrication in mind, unlike say, fabbers.org
[22:33:06] <kanzure> is fabbers.org the MIT CBA group?
[22:33:22] <fenn> uh, i mean fabathome.org
[22:33:32] <fenn> no, it's hod lipson (cornell)
[22:33:34] <kanzure> ah
[22:34:02] <kanzure> yeah, I know a few guys over at CBA, but not the fabathome.org group
[22:34:24] <kanzure> anyway,
[22:34:25] <fenn> who do you know at CBA?
[22:34:32] <kanzure> David Dalrymple
[22:34:40] <kanzure> the kid trained by Kurzweil :)
[22:35:05] <ds2> hmmmm there are other people here interested in reprap?
[22:35:13] <kanzure> ds2: of course
[22:35:27] <kanzure> so as far as I can tell, the trick is to include metal fabrication into the design of RepRap
[22:35:33] <kanzure> and also the transistor-production capabilities
[22:35:45] <kanzure> plus a way to load the resources and have it process the stuff on board to spit out a copy of itself
[22:35:47] <ds2> kanzure: do you have one working yet?
[22:35:54] <kanzure> ds2: No.
[22:35:58] <kanzure> But it looks relatively simple.
[22:36:04] <ds2> kanzure: do you plan to have one working in the near future?
[22:36:06] <fenn> i know this might sound weird, but have they considered vacuum tubes?
[22:36:16] <kanzure> fenn: vacuums are hard to generate, right?
[22:36:24] <kanzure> fenn: relays, however, are a good idea, but this too requires semiconductors
[22:36:26] <kanzure> fenn: check out my notes on this topic
[22:36:32] <kanzure> http://heybryan.org/alternate_transistors.html
[22:36:46] <kanzure> the LiquiFET idea was interesting, but it only works on the nanoscale since it runs off of capillary action
[22:36:47] <fenn> not especially. a 'sufficient' vacuum is much easier to accomplish with smaller sizes
[22:37:05] <kanzure> I am wondering how much circuitry they really need
[22:37:16] <kanzure> all programs can be expressed as an electronic circuit
[22:37:17] <fenn> anyway, if you're using a pre-made computer then you might as well throw in some real electronics
[22:37:18] <kanzure> instead of on a microprocessor
[22:37:30] <kanzure> for example, I've seen JPEG encoders made out of transistors ;)
[22:37:39] <ds2> fenn: is there another simple way besides mod'ing a compressor or using an old fridge/air conditioner?
[22:37:47] <kanzure> so if you're certain about the programming, might as well just run your code through a compiler and generate an output of simple circuit components
[22:37:59] <kanzure> a compressor is easy to build with metal working tools
[22:38:10] <kanzure> at least, that's what I thought yesterday
[22:38:15] <fenn> hmm. dont even need electronics at that point. just make mechanical devices that do what you want
[22:38:17] <kanzure> (was looking up ways to do LOX fabrication in the home for rockets)
[22:38:29] <fenn> check out thermoacousic coolers
[22:38:35] <kanzure> fenn: perhaps; the gear combinations would be rather difficult to conceptualize. I guess a computer could figure that out for us.
[22:38:41] <kanzure> actually, that's not a bad idea
[22:38:53] <kanzure> there's no reason we can't implement building instructions via gears and so on
[22:41:13] <fenn> here we go: http://www.kettering.edu/~drussell/Publications/ThermoDemo.pdf
[22:41:43] <kanzure> I need to run, but when I get back I think I'll do a review of DIY mills/lathes/CNC machines and then see if all of the parts could possibly be fabricated on a modified RepRap, and then look for anybody doing pseudocode -> mechanical relays
[22:42:37] <fenn> well, i'll be here (unless i'm not :))
[22:42:48] <kanzure> alright, cya
[22:48:46] <tomp2> can reprap make anything to any tolerance ( better than carpenter tolerance?) not being a smartass but the cylinder replications i saw looked like bad cake decorating & would never fit a piston. maybe there's been some major imrovement.
[22:53:40] <alex_joni> don't insult carpenters :P
[22:54:52] <tomp2> :)
[22:55:22] <BigJohnT> if I was a carpenter and i had a hammer
[22:57:00] <tomp2> i just looked at the site, and there were some parts that were not crap. some holes were actually roundish. and some surfaces flatish.
[22:57:35] <alex_joni> they might have been fixed afterwards
[22:57:46] <alex_joni> tomp2: this? http://reprap.org/bin/view/Main/ItemsMade
[22:58:30] <tomp2> http://picasaweb.google.co.uk/VikOlliver/RepRap02/photo#5120251089181296642
[22:58:44] <tomp2> still looked 'nervous'
[22:59:17] <tomp2> but way better than a couple years go
[22:59:20] <tomp2> ago
[22:59:48] <alex_joni> that's a part made by a 300$ machine vs. a part made by a prof. machine?
[23:00:21] <tomp2> i thought it was all done by that tinkertoy cat cage machine
[23:00:45] <tomp2> http://picasaweb.google.co.uk/VikOlliver/RepRap02/photo#5155848747460755458
[23:01:13] <fenn> tomp2: the parts with "precision" holes are usually made with a ground rod as a sort of temporary mold
[23:01:22] <tomp2> ah!
[23:01:38] <tomp2> cake decorating with a ruler... the cheaters!
[23:01:44] <fenn> but it has no metalworking or ceramic baking capability so i dont really get it
[23:02:43] <fenn> i mean it's a neat design for an easily fabricated goo-squirter, but it's so lacking in functionality needed for self replication that i dont know where to begin
[23:03:40] <tomp2> fenn: my first hit on 'portmanteaux' for an example turned up 'islamofacism' which my mind butchered into I Slam O something and i had to read the article to get it right.
[23:03:57] <fenn> one major advantage of a hexapod (besides the fact that it can actually cut metal) is that you can calibrate out any errors in the structure (except lash/play)
[23:04:23] <tomp2> omni-triangulated
[23:04:49] <fenn> so, provided you have a reference to calibrate it with, the 1000th generation will be as accurate as the second
[23:05:03] <fenn> (or first, depending how you define it)
[23:06:30] <tomp2> what kind of a reference,a form in it's work envelope that is probed to the same values?
[23:06:44] <fenn> i was thinking a double ball bar
[23:07:04] <alex_joni> hmmm... http://reprap.org/pub/Main/ItemsMade/vik-parts-made-25-jan-2008.jpg
[23:07:22] <fenn> or even just a cylinder on a flat plate (with an electrical contact probe)
[23:07:24] <tomp2> ball bars are nice tools,
[23:08:17] <tomp2> alex_joni: yeah they're not crap, but not precision either
[23:08:18] <fenn> if you use the hexapod in "lathe mode" you can make the cylinder, and then probe it to measure the new cylinder's dimensions, and voila you have a new reference
[23:08:50] <fenn> +/- your probe's uncertainty each generation
[23:08:56] <alex_joni> I think the chosen material is crap
[23:09:14] <alex_joni> syntherized ceramics should be way better
[23:09:26] <fenn> i think they chose goo because it's easily recycled
[23:09:37] <tomp2> didnt someone use sugar?
[23:09:40] <fenn> yes
[23:10:21] <tomp2> why didnt hexapod/stewart platform get more popular?
[23:10:25] <fenn> alex_joni: without machining (or at least wet-forming) then ceramic will have the same bad surface finish and accuracy issues as thermoplastic
[23:10:44] <fenn> tomp2: you're asking the wrong guy :)
[23:11:13] <fenn> tomp2: if you figure it out, please let me know
[23:12:30] <tomp2> :)
[23:13:12] <tomp2> yeah, uni michigan, grand rapids has a grant to make a machine shop for space. they want a little box that can grow parts, they plan to use layer laser sintering and then 'dry edm' to refine the form and surface. project must have 100x100x100mm cube volume capability, and better tan 5um precision (hah!)
[23:13:34] <cradek> if I had to guess (and I do, since I've done no research): it's hard to build. the software to run it is hard.
[23:13:55] <tomp2> 'start trek replicator' great idea