#emc | Logs for 2005-10-19

Back
[00:00:34] <RIO> RIO is now known as roby
[00:01:16] <roby> roby is now known as robydebian
[00:01:28] <Jacky^> doh !
[00:01:32] <Jacky^> * Jacky^ sighs
[00:14:44] <cncuser> can i strip down the options in rtai a little ?
[00:14:56] <cncuser> is there some overview on what is needed by emc ?
[00:15:06] <alex_joni> hmm.. modules?
[00:15:10] <alex_joni> or what do you mean?
[00:15:45] <cncuser> the modules and the aps
[00:16:21] <cncuser> i go for a 20mb iso, including axis :)
[00:17:21] <alex_joni> make it 32 ;D
[00:17:34] <alex_joni> aps you can definately scrap
[00:17:42] <alex_joni> maybe leave one for testing rtai
[00:18:09] <cncuser> yes, im still thinking about that. for automatically calculating of period
[00:18:23] <alex_joni> that's been done before..
[00:18:32] <alex_joni> ok.. now on the modules
[00:18:35] <cncuser> and did work ?
[00:18:45] <cncuser> where is the code ?
[00:18:51] <cncuser> .. /script
[00:18:58] <alex_joni> somewhere on the web :)
[00:19:05] <cncuser> hmm
[00:19:05] <alex_joni> can't recall
[00:19:38] <alex_joni> ok.. this is needed from rtai:
[00:19:39] <alex_joni> adeos rtai_hal rtai_ksched rtai_fifos rtai_shm rtai_sem rtai_math
[00:19:50] <cncuser> ok, thanx
[00:21:27] <alex_joni> * alex_joni looks for the code
[00:21:56] <rcsu> MANY THANKS to alex_joni !!!
[00:22:18] <alex_joni> rcsu: that's why we are here :P
[00:22:27] <alex_joni> sorry it didn't work fully :)
[00:22:29] <alex_joni> maybe tomorrow
[00:23:08] <rcsu> yes, it dont work, but the steps can solve it
[00:23:37] <rcsu> i have a really good feeling for that
[00:23:50] <rcsu> g8 all, time for bed
[00:24:56] <Jymmm> any thoughts or suggestions to deal with steppers skipping when ran for 2+ hours ?
[00:25:19] <Jacky^> Jymmm: im having the same issue ..
[00:25:29] <cncuser> nevermid alex_joni
[00:25:29] <Jacky^> after a lot it is working
[00:25:43] <anonimasu> Jymmm: do they get hot?
[00:25:49] <Jacky^> and my motors are very hot ..
[00:25:51] <anonimasu> um nevermind
[00:25:54] <anonimasu> they do get very hot..
[00:26:02] <anonimasu> machine heatsinks
[00:26:07] <Jymmm> anonimasu: I forogt to actually check, but I am suspecting thermal issues.
[00:26:09] <Jacky^> i wonder a crap ..
[00:26:24] <anonimasu> Jymmm: it's most likely a heat issue
[00:26:25] <Jacky^> what if an heatsink is used on the motor ?
[00:26:27] <Jymmm> anonimasu they are mounted to a big slab of aluminum
[00:26:30] <anonimasu> * anonimasu has felt it before..
[00:26:37] <Jacky^> may with a fan too ..
[00:26:38] <alex_joni> cncuser: why?
[00:27:14] <Jymmm> anonimasu they are being driven with 24vdc @ 3A, what if I up the voltage/amperage?
[00:27:22] <Jymmm> 48VDC instead
[00:27:37] <cncuser> alex_joni: i ment nevermind. its not taht important. i may just use a table with user submittet period / cpu mapping.
[00:27:40] <anonimasu> Jymmm: no idea..
[00:27:46] <anonimasu> Jymmm: I just know steppers get very hot..
[00:27:47] <Jacky^> Jymmm: the mine are suppling about 35 volts ..
[00:27:58] <Jacky^> i tried reducing current, same issue
[00:28:06] <Jacky^> seems not a current prob .
[00:28:09] <anonimasu> I am running mine at 48v and 200a..
[00:28:10] <Jacky^> for me
[00:28:19] <Jymmm> 20A ?! wth?
[00:28:28] <anonimasu> welder as power supply ;)
[00:28:49] <anonimasu> not that my machine takes 200A
[00:29:35] <anonimasu> Jymmm: could you check your motors to see if they are hot when you experience it agagin
[00:29:38] <anonimasu> again..
[00:30:04] <alex_joni> cncuser: http://people.freenet.de/LinuxCNC/files/emcconf.1.0.0a.tar
[00:30:11] <alex_joni> not sure it does what I said
[00:30:19] <Jymmm> anonimasu I'll todd on a thermal probe. this has happened twice so far. 1st time in 2 hours, yesterday at 3 hours.
[00:30:25] <Jymmm> s/todd/toss/
[00:30:27] <Jacky^> Jymmm: how much feedrate youre going ?
[00:30:34] <Jymmm> Jacky^ 100IPM
[00:30:38] <Jacky^> wow
[00:30:48] <anonimasu> you might be pushing it..
[00:30:48] <Jacky^> nah .. im about 30
[00:31:04] <Jymmm> anonimasu it's fine for the first two or three hours.
[00:31:15] <Jacky^> the mine up 30 always lost steps !
[00:31:16] <anonimasu> Jymmm: one step here and another there..
[00:31:25] <Jacky^> try to reduce a bit ..
[00:31:30] <alex_joni> here a step there a step
[00:31:35] <alex_joni> everywhere a fscking step
[00:31:39] <Jymmm> lol
[00:31:40] <anonimasu> Jymmm: that sounds like pushing it..
[00:32:02] <alex_joni> * alex_joni goes to bed
[00:32:02] <Jacky^> Jymmm: try to slow down a bit
[00:32:05] <alex_joni> errr...
[00:32:08] <alex_joni> * alex_joni steps to bed
[00:32:09] <Jacky^> i think you will sove
[00:32:12] <Jymmm> anonimasu I cna actually get 180IPM out of it, so going 100IPM does seem like I am backing off (to me).
[00:32:14] <anonimasu> night alex_joni
[00:32:16] <Jacky^> night alex_joni :)
[00:32:19] <alex_joni> night guys
[00:32:29] <Jymmm> night alex_joni
[00:32:30] <anonimasu> Jymmm: you are faster then I am cool :)
[00:32:40] <Jacky^> Jymmm: or .. maybe
[00:32:42] <Jymmm> anonimasu ha!
[00:32:55] <anonimasu> not for too long..
[00:32:57] <anonimasu> ;)
[00:33:01] <Jacky^> some axis is working aroung resonace frequency
[00:33:07] <anonimasu> about 120 is the limit with my setup..
[00:33:07] <Jacky^> around*
[00:33:09] <Jymmm> anonimasu well, if I HAVE to I'll buy geckos to get the 48VDC+ out of em
[00:33:40] <anonimasu> I am ordering a set of spare servos off surpluscentre soon
[00:33:41] <anonimasu> :)
[00:33:45] <Jymmm> I'll turn it down to 80 IPM
[00:33:54] <anonimasu> it's more likely the accels..
[00:33:56] <Jacky^> Jymmm: you know stepper have resonance frequency where can lost a lot ..
[00:34:00] <Jymmm> I bought twwo blanks today.
[00:34:16] <Jymmm> Jacky^ yeah, what about it now?
[00:34:24] <anonimasu> Jymmm: try limiting your accels a bit..
[00:34:28] <Jacky^> it could be work fine fro 80 to 90 ipm then from 120-140 and not to 100 !
[00:34:38] <Jacky^> try different feedrate
[00:35:00] <anonimasu> you might have some move where it tries to accel too fast and fscks
[00:35:05] <Jacky^> otherwise.. slow down :(
[00:35:10] <Jymmm> See, only thermal explains the lost steps after 3 hours. not the speed.
[00:35:59] <Jacky^> anonimasu: i'm having the same problem
[00:36:11] <Jacky^> but at very different speed 30 ipm
[00:36:30] <anonimasu> the only problem I've had with my stepper is that the accel's have been too fast
[00:36:39] <Jacky^> already tried all values for accell .. nothing to do
[00:36:51] <Jacky^> it tooke to me a mounth :/
[00:36:53] <cncuser> alex_joni: thank you very much. it does some calculations with bogomis and takes care of the cputype too i think :)
[00:37:10] <anonimasu> cncuser: he went to sleep
[00:37:31] <cncuser> ic
[00:38:48] <anonimasu> I hope I get my spindle back togther tomorrow night
[00:39:40] <anonimasu> * anonimasu sees projects piling up
[00:40:09] <anonimasu> got prototype stuff for work to make
[00:40:43] <anonimasu> and new bearing mounts
[00:40:55] <Jacky^> steppers hurt me :( i cand understand what mean lost the job after hours ..
[00:41:02] <Jacky^> damn it
[00:41:43] <anonimasu> well atleast you not making Ti scrap ;)
[00:41:53] <Jacky^> :/
[00:42:27] <Jacky^> i just decided to use this machine for 2d profiling only
[00:42:37] <Jacky^> not 3d relief
[00:42:40] <anonimasu> Jacky^: why not 3d?
[00:42:50] <Jacky^> because it take a lot of time
[00:42:55] <Jacky^> many hours
[00:43:08] <anonimasu> yeah
[00:43:09] <Jacky^> and all axis are stressed a lot
[00:43:24] <anonimasu> * anonimasu longs until he can do 3d contouring at 500ipm
[00:43:25] <Jymmm> Jacky^ prozace for steppers?
[00:43:30] <Jymmm> prozac
[00:43:45] <Jacky^> prozac ? what is ?
[00:43:53] <anonimasu> antidepressant..
[00:43:54] <Jacky^> let me check ..
[00:43:58] <Jacky^> hehehee
[00:44:00] <Jacky^> lol
[00:44:20] <Jacky^> Jymmm: i think servos..
[00:44:24] <Jacky^> or nothing
[00:44:26] <Jacky^> :)
[00:45:06] <cncuser> whats the thing with servos and steppers? is ther eone better for a job then the other. or is it the price ?
[00:45:30] <Jacky^> ouch ..
[00:45:40] <Jacky^> probably is the same discussion:
[00:45:48] <Jacky^> whats better parport or usb ?
[00:46:00] <Jacky^> it depend ..
[00:46:06] <anonimasu> depends on what you want to do
[00:46:18] <cncuser> usb has more bandwith. parport less latency i assume.
[00:46:25] <anonimasu> correctly sized steppers will work fine
[00:46:34] <dmess> i want to rock and roll on aircraft landing gear..
[00:46:50] <anonimasu> dmess: ? morning
[00:46:56] <dmess> gonna retro an old BMC-80 toshiba
[00:47:08] <anonimasu> dmess: I hate you :)
[00:47:15] <dmess> why??
[00:47:29] <anonimasu> because you can get your hands on good machines to retrofit
[00:47:32] <cncuser> anonimasi: hmmm
[00:47:37] <dmess> BTD-150??
[00:47:48] <anonimasu> dmess: I cant get any machines for a good price
[00:47:52] <anonimasu> a bp is 5000eur
[00:47:53] <dmess> you want one.. i'll ship it you buy it
[00:47:54] <cncuser> anonimasi: so why servos ?
[00:48:18] <anonimasu> cncuser: in general they provide more speed then steppers can
[00:48:19] <les> cncuser: servo cost/performance is better on on big machines...stepper on very small machines
[00:48:20] <Jacky^> speed
[00:48:24] <Jacky^> for me..
[00:48:36] <Jacky^> and power too
[00:48:39] <cncuser> ic
[00:48:40] <anonimasu> les: how's stuff going?
[00:48:51] <dmess> but i wanna know WHO can make it all work in my machine???
[00:49:02] <cncuser> howabout the servo feasture of reporting missed turns ?
[00:49:09] <cncuser> is it used ?
[00:49:12] <les> I am kinda crazy...had more technical breaktroughs
[00:49:16] <anonimasu> cncuser: you dont miss turns with a servo..
[00:49:25] <cncuser> anonimasu: haha
[00:49:32] <Jacky^> :)
[00:49:53] <anonimasu> usually the control compensates continously for lost steps/following error within reasonable limits..
[00:50:07] <anonimasu> a emco compact is like 3800eur :/
[00:50:07] <dmess> i'll be needin some input from you LES if you wouldnt mind.. ; )
[00:50:09] <cncuser> anonimasu: shure, its a matter of torque. that lets you miss a position. it must know it. or keep the wrong position
[00:50:11] <les> servo/stepper question also depends on the user needs
[00:50:16] <les> like jacky...
[00:50:21] <les> he should be servo.
[00:50:39] <anonimasu> or correctly sized steppers..
[00:51:18] <Jacky^> oversized would be better :P
[00:51:22] <cncuser> well, if i want to drill pcb s and aluminium
[00:51:24] <Jacky^> hehe.. kidding
[00:51:26] <cncuser> in postcardsize
[00:51:34] <les> If you want to have fun with your cnc...consider steppers
[00:51:46] <les> If you want to make money...use servos
[00:51:57] <cncuser> hahaha
[00:52:04] <anonimasu> ttp://195.149.144.113/images/52/52-52982916752.jpg
[00:52:21] <Jacky^> someone has been very lucky in ebay with servos
[00:52:42] <les> if you want to have fun and make money....use servos unless you are selling hobby stepper machines!
[00:52:44] <Jacky^> found 3 nice motors for $ 300
[00:52:55] <cncuser> les: i dont want to make money :) and use steppers. but the thingy i have here cant do aluminium (for its base don aluminum)
[00:52:56] <anonimasu> I dropped the guy a mail asking for the size..
[00:53:02] <Jacky^> when the price is 400-500 each :\
[00:53:27] <Jacky^> this thing wont happen to me !
[00:53:57] <les> cncuser: if sized right, and you don't want to make money...steppers can work.
[00:54:02] <les> with aluminum
[00:54:21] <cncuser> ok, thanks
[00:54:27] <dmess> and use cheaper tooling too
[00:54:28] <cncuser> well, gotta gpo to bad
[00:54:30] <cncuser> brd
[00:54:32] <cncuser> e
[00:54:33] <cncuser> ;)
[00:54:43] <Jacky^> cncuser: if you have time .. stay tuned on ebay, and look around could be cheaper
[00:54:50] <Jacky^> well
[00:55:52] <Jacky^> i'm right having this big problem now ..
[00:56:02] <Jacky^> where to buy the motors
[00:56:37] <dmess> goto the local stepper motor store...
[00:56:51] <Jacky^> car motor ?
[00:57:02] <Jacky^> :)
[00:57:12] <dmess> siemens.. fanuc...tosnuc....fagor....
[00:57:26] <dmess> heidenhain
[00:57:28] <Jacky^> yeah.. there are a lot
[00:57:55] <dmess> then there are the american home growns...
[00:58:12] <Jacky^> oh .. im in italy
[00:58:23] <Jacky^> id like to get it here
[00:58:31] <dmess> at least in CANADA our home grown IS the best
[00:58:34] <Jacky^> at least the motors
[00:58:54] <Jacky^> i'm buyng all from outside.. drivers, rails
[00:59:26] <dmess> fagor is spanish i believe
[00:59:28] <Jacky^> hard ship ..
[00:59:44] <Jacky^> there are a lot also here
[00:59:53] <Jacky^> but the prices are high :(
[00:59:58] <Jacky^> very high ..
[01:00:37] <dmess> any DMG machines in the area
[01:00:47] <Jacky^> but les have right when sayd ..
[01:00:58] <Jacky^> cnc are not cheaper
[01:01:27] <dmess> cheap and cnc.... do not go together
[01:01:37] <Jacky^> no
[01:01:42] <dmess> nope
[01:02:39] <Jacky^> anyway .. a commercial machine
[01:02:56] <Jacky^> cost euro 27K
[01:03:00] <dmess> unless you want a cnc skill saw.. with 1/8 accuracy and 6 ipm max speed
[01:06:28] <Jacky^> nah .. im more fast with the hands :)
[01:08:03] <Jacky^> who was looking for a pcb machine ?
[01:08:07] <Jacky^> http://www.cnc-dag50.it/index.php?set_albumName=album02&option=com_gallery&Itemid=26&include=view_album.php
[01:08:15] <Jacky^> thats nicee :)
[01:08:17] <dmess> precisely... cheap cnc... now.. do it by hand
[01:09:03] <Jacky^> nice toy :)
[01:10:11] <dmess> too mant toys filling up space
[01:10:17] <dmess> many
[01:11:00] <Jacky^> it seem to me cncuser was looking for a small machine to use for pbc
[01:12:20] <anonimasu> night guys
[01:12:27] <Jacky^> g night anonimasu
[01:21:28] <dmess> any mad scientists in the house??
[01:23:07] <les> well...I'm mad....will that work?
[01:25:13] <dmess> you'l do...
[01:25:19] <les> heh
[01:25:51] <dmess> piezo speakers... can we run them into the ultrasonic range
[01:26:36] <les> generally the cone resonance determines the hf reponce
[01:26:43] <les> which speakers?
[01:27:00] <dmess> and target frequencies to smash up certain elements
[01:27:19] <dmess> you tell ME which speakers..
[01:27:27] <les> ok
[01:27:39] <les> elements as in the priodic table?
[01:27:46] <les> periodic
[01:28:31] <dmess> yeah that one
[01:29:18] <Jacky^> dmess: ultrasonic ? to kill mouses ?
[01:29:30] <Jacky^> :P
[01:29:41] <Jacky^> i use a cat :)
[01:29:41] <dmess> the only frequency we dont wanna destroy is the diamonds
[01:29:43] <les> ok. Atomic resonances that selectively ionoze elements typically occur in the x-ray frequency range
[01:30:07] <les> for molecules resonances typically occur in the ir to uv range
[01:30:19] <dmess> so we pump it up a notch or two
[01:30:26] <Jacky^> ugh
[01:30:32] <Jacky^> g night guys
[01:30:37] <les> good night
[01:30:42] <Jacky^> Jacky^ is now known as Jacky^afk
[01:30:44] <dmess> gnite
[01:31:04] <les> energy = planck's constant times the wavenumber
[01:31:32] <dmess> where the hell am i supposed to find an x ray generator
[01:32:17] <les> oh....30kv electron beam...a tungsten target....enough to grow two heads
[01:32:31] <dmess> you got one???
[01:32:46] <dmess> free to a good home
[01:32:54] <les> nope...one head is enough for me
[01:33:18] <les> I have had xray soures though
[01:33:26] <les> sources
[01:33:31] <dmess> im a twin-less twin now... i had two before
[01:33:38] <dmess> its gooder
[01:33:38] <les> haha
[01:34:22] <les> I was in fact working with High intensity ultrasonics today
[01:34:33] <les> I have a research commission for that
[01:34:41] <dmess> i used microwaves last ite and created a noxious stinck in the kitchen..
[01:35:02] <les> much fun can be had with microwaves
[01:35:20] <dmess> grew a diamond encusted crucible though ; )
[01:35:27] <les> heh
[01:35:45] <dmess> i should have pis tomorrow nite.
[01:35:53] <dmess> pics
[01:36:15] <dmess> i hope i piss BEFORE tomorrow nite
[01:36:18] <les> oh, ok. that seemed like a long time to hold it
[01:38:45] <dmess> average paticle size i cooked was about .001" dia.... particulate cooked up with magic secret medicine man knowledge to form a black cubic substance
[01:38:56] <les> I have a can of beer here, and I hope to piss in about 5 minutes.
[01:39:18] <dmess> i do too and tommorow' too far away
[01:39:41] <les> what are you making? bucky balls? No, .001 is too big.
[01:40:28] <dmess> i added about 1/4 of graphite by volume to the ore prior
[01:40:43] <les> oh the kimberlite?
[01:40:58] <dmess> i have smaller dust.. but why process it yet
[01:41:18] <dmess> yes kimberlite mass samples
[01:41:38] <les> you need a bunch of caustic soda I think
[01:42:06] <dmess> i wanna staw away from caustic dissolution
[01:42:11] <dmess> stay
[01:42:29] <les> don't want to dissolve and stuff huh?
[01:42:36] <dmess> i have some sitting in phosoric acid as we speak...
[01:42:50] <les> hmm
[01:43:04] <dmess> phosphoric
[01:43:32] <dmess> id rather not.. need 1 ton per hr... cant dissolve that
[01:43:47] <les> low silica ore...lots of aluminum/magnesium minerals
[01:43:52] <les> garnet
[01:43:58] <les> olivine
[01:44:05] <les> stuff like that
[01:44:17] <les> and a little carbon...
[01:44:46] <dmess> not sure... did a mag separation.. got about 3-5 %
[01:45:22] <dmess> screened the rest.. sifted thru glass the carbon weaves...
[01:45:34] <les> I don't know if diamonds hang with the ferromagnetic stuff
[01:45:59] <dmess> no its a slush cull
[01:46:43] <les> sounds like the weeding out of the worst of my old girlfriends
[01:47:17] <les> no that is a slut cull
[01:47:33] <dmess> i never called HER my girlfriend... now you gotta die..
[01:47:40] <les> hahaha
[01:48:12] <dmess> i have a spot for ya.... remember
[01:48:24] <les> ha
[01:49:54] <dmess> ok so now i have som black flatish/cubic particulate occupying some of the same space as my diamonds...
[01:50:14] <les> graphitic?
[01:50:33] <dmess> i added graphite
[01:50:38] <les> ah
[01:50:57] <dmess> 1/4... its all like that or its diamonds
[01:51:19] <les> well impure diamonds might be black
[01:51:45] <dmess> how to extract the cooked off rock(silica)
[01:51:50] <les> or at least dark...depending on the dopant
[01:52:15] <les> silica? hmmm
[01:52:54] <dmess> well ore.. rock.. sand.. silicon... stop me if im lyin'
[01:53:06] <les> hydrofluoric acid?
[01:53:24] <dmess> i dont have any of THAT one..
[01:53:27] <les> kimbelite is supposed to be ultramafic...low silica?
[01:53:36] <les> kimberlite
[01:53:55] <dmess> and neither does work... they looked at me special when i asked too..
[01:54:23] <dmess> ok so what did i make in the nuker???
[01:54:41] <les> I have about 40 carats of industrial diamond in the shop
[01:54:43] <dmess> im still open to suggestions
[01:55:15] <dmess> from 3 old wheels... and the dresses layin around
[01:55:23] <les> I don't know
[01:55:26] <dmess> dressers
[01:55:44] <les> I am a lowly rocket scientist
[01:55:52] <les> at least by education.
[01:56:10] <les> dresses laying around? cool
[01:56:17] <dmess> and me a lowly methods engineer
[01:56:44] <dmess> we make STUFF
[01:57:14] <les> actually I make stuff too...only did aerospace in school
[01:57:18] <dmess> but i see DIAMONDS in my future
[01:57:22] <les> space shuttle.
[01:57:27] <les> a silly design.
[01:57:40] <dmess> cool glider.. ; )
[01:57:54] <les> so are rocks
[01:57:55] <dmess> i fly unpowered aircraft
[01:58:27] <les> I am a licsenced pilot
[01:58:38] <dmess> i remember..
[01:58:40] <les> flew gliders ony once
[01:58:50] <les> I called it engine failure.
[01:58:53] <dmess> not a hangy??
[01:59:07] <les> I have flown hangies yes
[01:59:15] <les> but this was not
[01:59:46] <dmess> oh ok.. in the bad old days of standards and home builts??
[02:00:37] <les> I built my own hangy as a teen
[02:00:48] <les> nearly killed myself in the thing
[02:01:08] <dmess> guts is good... makes ya dream
[02:01:24] <LawrenceG> les... dont use steel pipe next time....
[02:01:30] <les> haha
[02:01:40] <LawrenceG> water pipe is for water
[02:01:44] <dmess> near death experiences seem to have a similar effect
[02:01:47] <les> it was fibreglass wrapped bamboo
[02:01:58] <LawrenceG> hi tech composites
[02:02:03] <les> kinda
[02:02:11] <dmess> in '69
[02:02:26] <les> well...let's see
[02:02:29] <les> um
[02:02:38] <les> about "72
[02:02:45] <dmess> you must have had 30 -40 lbs a glass on there
[02:02:48] <les> I have pictures but not scanned
[02:02:56] <dmess> please... ; )
[02:03:05] <les> total weight was about 50 lbs
[02:03:27] <dmess> any kingpost or top wires??
[02:03:58] <les> no kingpost...but lots of wires. It was a rogallo
[02:04:53] <dmess> oh i quiver when i see them on the web for sal - in good used condition...
[02:04:58] <les> kind of a tringle shaped parachute really
[02:05:02] <les> but
[02:05:15] <les> two guys could pull it up with a rope
[02:05:38] <les> I stopped when I spun the thing...over water
[02:05:44] <dmess> 1 can pull me on a paraglider
[02:05:47] <les> It was a long time ago.
[02:06:02] <les> All I remember was that it hurt...
[02:06:04] <les> a lot.
[02:06:06] <dmess> over water was a good thing
[02:06:10] <les> yeah
[02:06:52] <dmess> i hit in from about 150' in the PG into the sod.... that hurt ALOT
[02:07:45] <les> Well I had not been to engineering school yet...so I did not know the principles of similitude.
[02:08:00] <les> I just scaled up 1/4 models
[02:08:18] <les> I did not know that I had to scale up the air too
[02:08:45] <dmess> a couple broken wrists... another break in a back wit 3 others... and 3 floating ribs that just wont stay home
[02:08:54] <les> yikes ouch
[02:09:22] <dmess> it sucks to be me some days..
[02:10:41] <dmess> so you think we'de need to get into the x-ray realm as opposed to ultrasonic
[02:13:14] <les> to separate elements yes
[02:13:27] <les> I did that in scoll
[02:13:38] <les> school
[02:14:00] <les> for Lawrence Livermore National Laboratory
[02:14:14] <les> which I cannot talk about much
[02:15:07] <dmess> still bound by promises eh..
[02:15:08] <les> but
[02:15:13] <les> in general
[02:15:35] <les> atomic resonances occur at x-ray frequencies
[02:16:10] <dmess> if hit with said res.. whats left??
[02:16:16] <les> we use x-ray spectroscopy to do atomic number contrast in electron microcopes for example
[02:16:33] <dmess> yup
[02:17:06] <les> well... you can selectively ionise elements...then separate with electrostatic forces
[02:17:55] <dmess> i thought but carbon and diamond will be similarilly charged.. wont they
[02:17:59] <les> diamond vs silicon I don't know...
[02:18:45] <les> yes I think the x-ray absorption spectra of graphite and diamond might be pretty similar
[02:19:20] <dmess> and i know ive added graphite to allow the cook off
[02:19:31] <les> hmm
[02:20:11] <dmess> but its the cooking of the other ore and what its turned into i need more info on
[02:20:43] <dmess> im sure somehow i made some glass in there too
[02:20:49] <les> I just don't know
[02:21:28] <les> to experiment you need a few more tons of high grade kimberlite
[02:21:30] <les> heh
[02:22:50] <dmess> well thanks for you time and knowledge les... but if i cant get the $$$ out of it the wife bitches about tons of stuff that "gonna be worth $$ some day"
[02:23:05] <les> yeah
[02:23:18] <dmess> they are funny like that
[02:23:26] <dmess> go figure
[02:23:28] <les> I don't have that problem ...anymore
[02:23:56] <dmess> brb
[02:23:59] <les> k
[02:28:47] <LawrenceG> hey les
[02:29:02] <les> hi lawrence
[02:29:26] <LawrenceG> I have been trying to think outside the box on this 3 phase motor drive stuff
[02:29:39] <les> still kinda nuts...had another step in the right direction with the air electricity generator
[02:30:08] <LawrenceG> there is no problem generating the drive...
[02:30:19] <les> My 0-1000 Hz 6.1 kW inverter came intoday
[02:30:57] <LawrenceG> but if thought of as a brushed dc motor where the drive signal just gives more torque (what I would like to emulate)
[02:31:39] <LawrenceG> or a stepper where you have a position to go to and the full drive current is on the windings all the time
[02:31:45] <les> I think in general current = k* torque still applies
[02:32:35] <LawrenceG> so some how, I need the drive signal to translate into a phase + current adjustment
[02:32:55] <LawrenceG> so that a high drive signal yeilds high speed + full torque
[02:33:11] <les> well I suspect phase is constant but current per leg is not
[02:33:11] <LawrenceG> and a low drive signal yields low torque and low speed
[02:33:32] <les> low current= low torque
[02:33:49] <les> low phase frequency= low speed
[02:34:04] <les> like a 3 pase ac syncronous motor
[02:34:07] <les> oops
[02:34:10] <les> spellig
[02:34:14] <les> hahaha
[02:34:27] <les> switching to my computer glasses
[02:34:30] <LawrenceG> true but you can have full torque at zero speed like a stepper
[02:34:40] <LawrenceG> because this is a pm rotor
[02:34:57] <les> I would think so
[02:35:26] <LawrenceG> with 16usteps one gets 192steps/rev very similar to a full step stepper
[02:35:54] <les> yeaha typical 1.8 degree /200 step
[02:36:56] <LawrenceG> using the drive command from the (emc) controller, one can just increase or decrease phase currents, but it wont move unless one advances or retards the phase at some rate
[02:37:11] <les> rutex microstepping drive right?
[02:38:07] <LawrenceG> I am not sure what they do.... I am looking for simple drive to home brew for these motors that lets emc control the servo loops
[02:38:42] <les> I am not sure really. Since "ac brushless" and "dc brushless" depend on the manufacturer's whims...
[02:39:04] <LawrenceG> a lot of the app notes run the drive in velocity mode where the drive has its own tuning params and servo loops to maintain a velocity
[02:39:05] <les> docs or a physical sanyo drive would help a bunch
[02:39:35] <les> oh...emc will work fine with a velocity drive
[02:39:44] <LawrenceG> my drives shipped by mule train today, so I may see them in 3 weeks or so
[02:39:45] <les> just don't use d...
[02:40:17] <les> that would be a potentially unsable fourth order system
[02:40:24] <les> unstable
[02:40:29] <LawrenceG> I dont really want the complexity of having servo loops in the drive and emc.... should be able to make it simpler
[02:40:40] <les> seems I need new computer glasses too
[02:41:25] <LawrenceG> emulate a dc brush motor.... more gas => more torque and speed
[02:41:28] <dmess> Les.. if i use butter for the grease table i should be able to float and separate the crystals. right
[02:41:52] <les> yes...I prefer torque drive....velocity amps were generally used in the distant past to offload pid calcs from the very slow processors
[02:41:56] <dmess> and dip the lobster in the left overs
[02:42:48] <les> dmess: I have read about using oily tables to separate diamonds due to their flat faceted surfaces
[02:45:13] <les> and oleophillic nature
[02:47:37] <les> what is butter? I don't know. Probably esters of oleic, linoleic, and linoleic fatty acids
[02:51:36] <les> glycolic esters.
[02:51:54] <les> specifically triglycerides.
[02:54:51] <Jymmm> hey les
[02:54:58] <les> hi jymmm
[02:55:07] <Jymmm> les that wood I bought, isn't papuk
[02:55:16] <les> what is it?
[02:55:26] <Jymmm> gettign spelling...
[02:55:32] <les> coco bolo?
[02:55:33] <Jymmm> Sepele
[02:55:39] <les> ah ok
[02:56:04] <Jymmm> oh not cocobolo.... that's WAY too expensive to play around with.
[02:56:13] <les> yeah
[02:56:18] <les> I have some
[02:56:18] <Jymmm> they have stuff that's $38 BF
[02:56:37] <les> only sepele I have is veneer samples
[02:56:53] <Jymmm> ah, these are 2" x 8" x .5"
[02:56:59] <Jymmm> $1.50/ea
[02:57:26] <les> my samples are old and faded...turned kinda brown
[02:57:36] <Jymmm> ah, bummer
[02:57:46] <les> UV does that
[02:57:54] <Jymmm> bloodwood looks cool,
[02:58:10] <les> even purpleheart turns brown after a while
[02:58:41] <Jymmm> I originally like purpleheart, but it lost it's luster
[02:58:52] <Jymmm> African Mahogany is neet too
[02:59:13] <les> for some reason bright color+ UVA photon= brown
[02:59:53] <Jymmm> that reminds me.... you seen any PU with UV inhibitors in it?
[02:59:56] <les> ah..african mahogany isn't really a true mahogany...it is Kyha
[02:59:59] <les> but
[03:00:05] <les> it looks nice.
[03:00:11] <Jymmm> yeah it does.
[03:00:37] <les> there are only two true mahoganies
[03:00:50] <les> swietenia mahogani
[03:00:52] <les> and
[03:00:52] <Jymmm> I got a pint of glow-in-the-dark paint today too =) First time I've seen it in a can.
[03:01:13] <les> macrophilla Mohogani
[03:01:54] <les> In chicago I did a whole room in glow in the dark paint
[03:01:59] <les> I was kinda
[03:02:02] <les> nuts.
[03:02:09] <les> stil am.
[03:02:12] <les> still
[03:02:14] <les> haha
[03:02:15] <Jymmm> I have 6 four foot blakc lights =)
[03:02:37] <Jymmm> from when I did up my place a few years ago
[03:02:42] <les> oh...you're nuts too...good
[03:03:07] <les> love black light.
[03:03:12] <Jymmm> redwood bark + day glow orange paint + blacklight == hot orange coals
[03:04:06] <les> sofa+ murine eye medicine in streaks= cool invisible black light designs
[03:04:22] <Jymmm> lol
[03:04:24] <les> == I mean
[03:04:26] <les> haha
[03:05:20] <Jymmm> just toss on the black caulderon on top of the bark and a lil (LOT) dry ice and looks really good!
[03:05:32] <les> wha
[03:05:35] <les> haha
[03:06:01] <Jymmm> and perfectly safe
[03:06:10] <les> we had fun and games with LN2 at intersil
[03:06:38] <les> had to throw away several gallons from the sorption pumps each shift
[03:06:45] <Jymmm> the bark absorbs just enough of the paint to make it look exactly like hot charcoals under the blacklight
[03:07:14] <les> I poured some down a girl's shirt...she poured some down my pants
[03:07:23] <Jymmm> LN2 ?
[03:07:27] <les> lots of smoke.
[03:07:31] <les> yeah
[03:07:32] <Jymmm> lol
[03:08:07] <Jymmm> oh LN2... duh.... I was thinking LOX
[03:08:15] <les> uh...don't try this at home kids.
[03:08:48] <Jymmm> could you imagine.... helium + LN2
[03:09:05] <Jymmm> floating lil smoke balls
[03:09:15] <les> if it had been LOX she wouldn't have tits and I wouldn't have kids
[03:09:42] <Jymmm> thats to say she had any tits to begin with =)
[03:09:55] <les> heh
[03:10:10] <les> she did. I know.
[03:10:16] <Jymmm> no orange paint anywhere today...
[03:10:18] <Jymmm> lol
[03:10:29] <Jymmm> les the bikini inspector #69
[03:10:37] <les> cough
[03:11:11] <Jymmm> rotf
[03:11:21] <les> K'zan is going to show up to bug us now....
[03:11:30] <Jymmm> hey, did you say you have/use some alcohol based stains?
[03:12:11] <les> I have some...I use them. They are metallic salt based
[03:12:19] <Jymmm> wood + PU coating + route sign + smear with water/alcohol stain + dry + PU again.
[03:12:38] <les> can be seen on the camo green turkey call on the website gallery picture
[03:13:24] <Jymmm> ah, seems overboard for my needs =)
[03:13:32] <les> PU+ alcohol dye= not miscible
[03:13:33] <les> but
[03:13:41] <Jymmm> just somethign I can rub in the v carvings
[03:13:41] <les> perhaps neat?
[03:13:58] <Jymmm> not what?
[03:14:05] <les> I use (gulp)
[03:14:12] <les> latex wall paint
[03:14:34] <Jymmm> I need something like a stain... maybe waterbased
[03:14:46] <Jymmm> nice read oak wood
[03:14:48] <Jymmm> red
[03:15:01] <Jymmm> just to accent the carving is all
[03:15:16] <les> well try the alcohol stuff
[03:15:20] <les> I have not
[03:15:23] <Jymmm> personally, I HATE paint of natural wood.
[03:15:28] <Jymmm> s/of/on/
[03:15:35] <Jymmm> I even hate stain on wood
[03:15:36] <les> well try it
[03:15:42] <les> perhaps a new look
[03:15:52] <les> tell me what happens
[03:15:52] <Jymmm> I looked today, didn't see any alcohol stain
[03:15:57] <Jymmm> only oil and water
[03:16:15] <Jymmm> very little water based stain too
[03:16:16] <les> woodworker's supply
[03:16:27] <les> google it
[03:16:31] <Jymmm> yeah, but a pint would last me years.
[03:16:45] <les> I buy dry powders
[03:16:52] <Jymmm> just dip in a rag, spear into the carving, and let dry.
[03:17:05] <Jymmm> you mix the powder with alcohol?
[03:17:16] <Jymmm> like denatured alcohol?
[03:17:26] <les> right...or alcohol...or shellac...or nitro laquer
[03:18:01] <les> they have oil, water, or alcohol powders
[03:18:04] <Jymmm> I always have denatured alcohol around, so a dry poweder would work. Do you have to weight out the powder?
[03:18:28] <les> nah, just put it in till it looks right
[03:18:48] <Jymmm> Yeah, so the more powder, the darker the color?
[03:19:06] <les> until it is a saturated solution yeah
[03:20:03] <Jymmm> does the alcohol based "bleed" / absorb into the wood more than water/oil based, or is it more of a way to suspend the powder?
[03:20:07] <les> it is a dye...not a pigment particle. It works on the molecular level, so it has clarity.
[03:20:20] <Jymmm> ah, very cool.
[03:20:31] <Jymmm> so the grain will show thru
[03:20:38] <les> the powder is not suspended...it is disolved
[03:20:50] <les> molecular level
[03:20:52] <Jymmm> kool aid!
[03:20:57] <les> yeah
[03:21:05] <les> exactly
[03:21:51] <Jymmm> so since you mix what you want, do they have multiple colors, or more of a yel/red/blu sorta thing?
[03:22:01] <Jymmm> brb
[03:22:22] <les> color mixtures tend to separate dur to "paper chromatographic" effects
[03:22:29] <les> due
[03:23:07] <LawrenceG> Jymmm, try some rit dyes from the grocery store... lots of colors an cheap,
[03:23:19] <les> yes lawrence
[03:23:37] <les> rit dye is water soluble mettallic salt dye
[03:24:18] <LawrenceG> not sure how it would like alcohol....
[03:24:51] <les> often they will dissolve in other solvents...but the colors are different
[03:29:56] <Jymmm> rit sounds good... are they a gel or concentrated liquid our of the bottle?
[03:31:27] <Jymmm> I like the idea of alcohol based colors... evaporates quickly.
[03:31:46] <Jymmm> so rit dyes + alcohol sounds right up my alley.
[03:32:00] <Jymmm> I do have a electronics scale that reads grams too
[03:32:07] <Jymmm> up to 25 pounds
[03:32:17] <Jymmm> not a druggy scale =)
[03:32:57] <Jymmm> les: Oh I found out that charcoal lighter fluid dissolves lithium grease very well btw.
[03:33:02] <LawrenceG> hmmmm 10 kilos...
[03:33:43] <les> sure any non polar aliphatic or aromatic hydrocarbon will
[03:33:53] <les> I like kerosene
[03:33:59] <Jymmm> les: You want to make some machines... the clickable links are wrong, but just change the numbers on the link http://www.korumdesign.com/cnc_mill_inde.html
[03:34:10] <Jymmm> les they have unscented charcoal fluid now
[03:35:14] <les> yes that is high flash, aromatic free aliphatic hydrocarbon
[03:36:52] <Jymmm> http://www.korumdesign.com/cncmill19.html
[03:37:02] <Jymmm> the numbers go from 1 to 28
[03:37:23] <Jymmm> I want one of these blocks he's using, but a electronic version
[03:37:45] <les> looking
[03:39:17] <Jymmm> http://www.korumdesign.com/cncmill8.html
[03:39:34] <Jymmm> http://www.korumdesign.com/cncmill9.html
[03:40:03] <Jymmm> http://www.korumdesign.com/cncmill10.html
[03:40:36] <les> would be nice
[03:40:38] <Jymmm> you could make those les
[03:40:43] <les> hmm
[03:40:48] <Jymmm> he's in italy
[03:42:03] <Jymmm> les http://www.korumdesign.com/cncmill20.html
[03:42:05] <les> ohI Ideas
[03:42:23] <les> I used to do that cad/cam progrm
[03:42:31] <Jymmm> what is it?
[03:42:46] <les> just a high end cad cam
[03:42:50] <Jymmm> ah, ok.
[03:43:13] <Jymmm> heh, he has my kybd =)
[03:43:23] <les> cost a gazillion bucks+ more gazillions maint fee
[03:43:41] <Jymmm> oh is that all =)
[03:43:48] <les> heh yeah
[03:44:46] <Jymmm> les I was talking to the manager of the mill at the lumber yard today.... they've been thiking of getting a CNC machine JSUT to do picket fencing
[03:45:03] <les> really
[03:45:08] <Jymmm> seriously
[03:45:29] <Jymmm> thats all they wanted it for... just picket fencing. they're doing it now on the bandsaw
[03:45:48] <les> hmm
[03:46:10] <LawrenceG> cool lots of different pointy ends
[03:46:15] <LawrenceG> and fluting
[03:46:16] <Jymmm> lol
[03:46:51] <LawrenceG> how many can you cut from a sheet of mdf? lasts until first rain
[03:47:02] <les> well whatever makes money I guess
[03:47:12] <Jymmm> no, they do it from stock lumber, not mdf
[03:47:12] <les> haha wet sawdust
[03:47:21] <les> well there is always medex
[03:47:23] <Jymmm> stinky wet sawdust
[03:47:29] <Jymmm> medex?
[03:47:29] <les> waterproof MDF
[03:47:38] <Jymmm> HazMat too?
[03:47:48] <les> big time sign stuff Jymmm
[03:47:51] <les> google it
[03:48:14] <Jymmm> travel insurance
[03:48:15] <Jymmm> =)
[03:52:31] <les> I might use it too...not doing many signs right now...too busy with engineering
[03:52:33] <les> but
[03:53:11] <les> I did order a 4x8x1.5" sheet of 15 lb urethane foam
[03:53:21] <les> yesterday.
[03:53:29] <les> $300!!!!!
[03:53:34] <Jymmm> lol, I though you orded last week?
[03:53:35] <les> "Katrina"
[03:53:59] <les> yes I think I did actually
[03:54:10] <les> delivery is friday anyway
[03:54:15] <les> but
[03:54:20] <les> too much!!!
[03:54:21] <Jymmm> so you ordered last week and yesterday huh?
[03:54:32] <les> or not and haha
[03:54:39] <Jymmm> lol
[03:54:47] <les> I am ordering a lot of stuff lately
[03:55:04] <les> I get confused easily
[03:55:24] <Jymmm> well order me a cheeseburger, fries and a coke already
[03:55:38] <les> I have allotted $15K for shop improvements
[03:55:46] <les> and cheeseburgers
[03:55:48] <les> so
[03:55:58] <les> the UPS man knows me well
[03:56:27] <les> the big stuff is being delivered to the appple packing house loading dock nearby
[03:56:52] <les> I then pick it up and take it to the shop with the tractor
[03:57:11] <Jymmm> you DO have your pitchfork and straw hat dont you les?
[03:57:17] <les> today only one thing came
[03:57:21] <les> that vfd
[03:57:32] <Jymmm> "GREEEEEEEEEEEEEEEEEEEEEN ACRES is the place to be...."
[03:57:40] <les> pretty much
[03:58:17] <les> I never knew how absolutely accurate green acres was
[03:58:24] <Jymmm> now I'm excited to dry the rit dyes + alcohol
[03:58:38] <Jymmm> heh, ok Zaza
[03:58:41] <Jymmm> (sp)
[03:59:09] <les> who knows what will happen? Water and alcohol are both polar solvents.
[03:59:17] <les> psycodelic
[03:59:38] <Jymmm> sounds like it's on a controlled substance list somewhere =)
[04:00:04] <les> aw there is an MSDS sheet for WATER!!!!
[04:00:06] <Jymmm> what to you use as a base for your dyes?
[04:00:07] <les> really
[04:00:24] <Jymmm> I believe it... but is there a CAS # ?
[04:00:43] <les> I have oil, alcohol, and water based dyes
[04:00:54] <Jymmm> the salt ones
[04:01:01] <les> CAS for water...haha...probably
[04:01:10] <les> yeah the salt ones
[04:01:31] <Jymmm> Yeah, I learned about CAS #'s from a scientic freind of mine
[04:01:32] <les> commonly called aniline dye
[04:01:38] <les> but they are not
[04:02:35] <les> the water based dyes are the most light fast
[04:03:00] <les> get a kit from woodcrafter's supply....fun!
[04:03:18] <les> I also got my anodising dyes
[04:03:18] <Jymmm> Kinds funny... http://www.woodweb.com/knowledge_base/MDF_versus_particleboard.html
[04:03:21] <les> all colors
[04:03:26] <les> mil spec
[04:03:29] <Jymmm> oh, hows is that ?
[04:03:33] <les> prob similar
[04:03:48] <Jymmm> have you played with your anozdizing kit yet?
[04:03:58] <les> not yet
[04:04:05] <les> will set it up soon
[04:04:25] <Jymmm> well shit... you kepp buying all the toys, and I'll come over and play with them
[04:04:39] <Jymmm> if you wont
[04:04:59] <les> The toys are coming in faster than I can play with them I guess
[04:05:15] <les> that anodizing is gonna be cool I think though
[04:05:32] <les> you saw the pictures
[04:06:59] <les> All my product prototypes will be splash/gradient purple anodize soon
[04:07:01] <les> haha
[04:09:08] <les> well about time for bed for me
[04:09:16] <les> I sure had a great day
[04:09:23] <les> I am awfully happy
[04:11:06] <les> after literally starving for a while...
[04:11:24] <les> there is something making great money in my shop...
[04:11:31] <Jymmm> lol... well have a g'night les
[04:11:37] <les> and after what I saw today...
[04:11:45] <les> it will make a lot more.
[04:11:52] <les> good night
[04:11:54] <Jymmm> I wish I was making money insted of gettin 3 hours into a job and loose steps
[04:12:05] <Jymmm> twice
[04:12:21] <les> We need to talk about that step thing and get it fixed
[04:12:33] <les> tomorrow
[04:12:36] <Jymmm> tomorrow mornin
[04:12:40] <les> k
[04:12:40] <Jymmm> =)
[04:13:09] <les> goodnight
[04:13:19] <Jymmm> =)
[04:13:30] <les> night
[04:13:55] <les> les is now known as les_away
[04:18:16] <rayh> rayh is now known as rayh-away
[06:14:15] <CIA-14> 03paul_c * 10emc2-auto/wiki/ (13 files in 10 dirs): "Auto update wiki from a cron job. Wed Oct 19 05:30:01 BST 2005 "
[11:20:03] <Jacky^afk> Jacky^afk is now known as Jacky^
[12:08:53] <cncuser> good morning folks
[12:09:14] <Jacky^> hello cncuser
[12:09:20] <cncuser> hi jacky
[12:09:47] <Jacky^> was you looking for inexpensive cnc pbc machine ?
[12:09:59] <cncuser> jacky: yes :)
[12:10:17] <Jacky^> found some interesting pic ..
[12:10:20] <Jacky^> a moment
[12:10:23] <cncuser> :)
[12:11:28] <Jacky^> http://www.cnc-dag50.it/index.php?set_albumName=album02&option=com_gallery&Itemid=26&include=view_album.php
[12:11:40] <Jacky^> theres no plain, just the photos
[12:12:04] <Jacky^> but i could try to ask him for plain .. if youre interesting
[12:12:07] <Jacky^> :)
[12:12:10] <cncuser> hehe
[12:12:27] <cncuser> thank you, looks nice
[12:12:51] <Jacky^> :)
[12:12:55] <Jacky^> yeah
[12:13:42] <cncuser> hmm, i have some epson dfx8000 and one 5000+ i think im going to "mod" them someday :)
[12:14:05] <Jacky^> printers ?
[12:14:18] <cncuser> yes, heavy duty printers
[12:14:23] <Jacky^> good :)
[12:15:04] <cncuser> around 10 years old i think. heavy big monsters
[12:15:13] <Jacky^> hehe
[12:15:28] <cncuser> seems that he used printers too
[12:15:44] <Jacky^> uhmm no idea..
[12:15:54] <Jacky^> but seems simple to build
[12:16:04] <cncuser> at leas the sliding axis and the rubberband looks familiar
[12:16:07] <Jacky^> maybe the Z axis is a bit difficult ..
[12:17:24] <cncuser> i like the kitchen pic :)))
[12:17:33] <Jacky^> haha
[12:17:40] <Jacky^> he used strap ..
[12:17:49] <Jacky^> i think its fassssstt
[12:17:52] <cncuser> ehats that ?
[12:18:13] <Jacky^> instead using screw could be slow ..
[12:18:23] <cncuser> ok, the spindle is labeld canon...
[12:18:28] <Jacky^> just like a printer
[12:18:33] <Jacky^> :)
[12:19:01] <cncuser> hahahaha
[12:19:07] <cncuser> ok now i see
[12:19:24] <cncuser> he did exactly what my first thought bout building such a thin was.
[12:19:38] <Jacky^> :)
[12:19:50] <cncuser> put on printer 90? rotatet on the head of another printer :)
[12:20:17] <cncuser> the dfx8000 could drive a smaller matrix printer easily
[12:21:11] <Jacky^> here's another nice article http://www.fullnet.com/u/tomg/gooteecn.htm
[12:21:14] <cncuser> now if i put the pcb upside down i wouldnt have to change much.
[12:21:20] <Jacky^> probably you already know it ..
[12:22:19] <cncuser> no, looks nice
[12:22:22] <cncuser> trashy :))
[12:22:25] <Jacky^> hehe
[12:23:59] <Jacky^> with the straps should be very fast
[12:24:11] <Jacky^> the issue could be on the drivers
[12:24:25] <cncuser> http://www.argecy.com/DFX8000.html
[12:24:27] <Jacky^> would be nice to use microstep,
[12:24:34] <cncuser> 1066 characters per second
[12:24:39] <Jacky^> not full or half step ..
[12:24:54] <Jacky^> if youve the drivers its ok :)
[12:25:13] <cncuser> hmm, with drivers you mean ?
[12:25:33] <Jacky^> i think half step could not smooth enough for that applications
[12:25:38] <cncuser> the printers have steppers i assume (for paper transport)
[12:25:47] <cncuser> on the other axis... hmm, i got to open it up :)
[12:25:47] <Jacky^> i mean electronics controller
[12:26:02] <cncuser> the printer has a controller in it allready
[12:26:30] <Jacky^> usually, they use microstep to get smoothnees travel
[12:26:50] <cncuser> ic
[12:26:59] <cncuser> microstep is complicatet to do ?
[12:27:00] <Jacky^> otherwise the machine could be very.. nervous :)
[12:27:06] <cncuser> why not use gears ?
[12:27:10] <Jacky^> not complicated
[12:27:24] <Jacky^> but a little expensive, maybe..
[12:27:50] <Jacky^> try first with the drivers you already have
[12:28:42] <Jacky^> microstep is nice for precision and smoothness drive
[12:29:16] <Jacky^> that why many printers are using it
[12:30:16] <cncuser> ok
[12:30:20] <Jacky^> i tried to drive some printer carriage with his driver
[12:30:25] <Jacky^> using full step
[12:30:33] <Jacky^> and the response id hard ..
[12:30:37] <Jacky^> is
[12:31:16] <cncuser> well, i think i open up one of the babys right now and take a look at the interieur
[12:34:06] <cncuser> damnd, forget how heavy it is
[12:34:29] <Jacky^> hehe :)
[12:35:20] <cncuser> hahaha, theres a sticker on the back showing four hands carrying the thing :)
[12:36:57] <cncuser> i need one of thos exoscelet robos shown in the second alien movie :)
[12:37:08] <Jacky^> hehe
[12:37:43] <cncuser> im not joking :) i carried it from 20 meter to here and am still exhausted
[12:37:59] <Jacky^> :D
[12:38:39] <cncuser> also voicechat would be nice. i cant feel my arms ;)
[12:38:46] <cncuser> ok, enough
[12:58:36] <alex_joni_away> alex_joni_away is now known as alex_joni
[13:26:47] <cncuser> hmm
[13:27:01] <cncuser> hi alex joni
[13:27:06] <cncuser> bye
[13:31:10] <alex_joni> hi, bye
[13:31:13] <alex_joni> :D
[13:32:38] <Jacky^> Jacky^ is now known as Jacky^afk
[14:37:51] <les> For your viewing pleasure I made another short comment on G64 on the list.
[14:38:20] <les> but now, off to the shop
[14:48:28] <Jacky^afk> Jacky^afk is now known as Jacky^
[15:48:43] <alex_joni> les: user list?
[15:54:21] <fenn> anyone ever used cam.py with emc?
[15:54:38] <fenn> i wanna turn .svg files into g-code
[15:54:48] <alex_joni> not me
[15:55:20] <fenn> i'm wondering how hard it would be to write a translator that outputs g-code with sane accel/decel
[15:55:27] <fenn> from svg
[15:55:39] <alex_joni> no idea ;)
[15:55:47] <fenn> svg looks pretty easy to parse
[15:57:08] <jepler> what's cam.py?
[15:57:19] <fenn> a script they use at MIT
[15:57:23] <jepler> http://web.media.mit.edu/~neilg/fab/dist/cam.py ?
[15:57:26] <fenn> yeah
[15:57:38] <fenn> mostly for circuit boards it seems
[15:58:03] <jepler> * jepler glances at it
[15:58:06] <jepler> there's a lot of code in there
[15:58:15] <jepler> (and a gui?)
[15:58:18] <fenn> but it says "1D path following, 2D contour and raster, 3D slicing" so i wonder how the 3d slicing is
[15:58:42] <fenn> very minimal gui
[16:06:58] <alex_joni> jepler: http://fab.cba.mit.edu/tools/machining/nc/circuit-modela.html
[16:09:10] <les> well such things can be controlled from cam somewhat...very simply required accel = programmed velocity^2/ local radius of curvature
[16:10:21] <jepler> alex_joni: ah, chris and I use eagle for PCBs, but we have our own eagle to g-code converter
[16:10:35] <alex_joni> jepler: I know
[16:10:46] <les> just saw jmk's reply about emc2 having the TP bugfix....I checked...it does. But it is not line for line the same as TP.c in emc1.
[16:10:48] <alex_joni> just pointed you a link about cam.py
[16:10:53] <les> So I don't know...
[16:11:46] <jepler> I know I'm snarking, but somebody should tell the cam.py author that creating a GUI is different from just making a button or entry field for each tunable parameter a program has
[16:12:56] <alex_joni> les: just try it out.. :)
[16:13:23] <alex_joni> you'll see that it works better than emc1 .. and you'll never go back :)
[16:14:01] <alex_joni> jepler: b&h are on vacation :(
[16:14:10] <fenn> alex_joni: yeah, after you patch your kernel for rtai support
[16:14:27] <alex_joni> fenn: and emc doesn't need a patched kernel?
[16:14:32] <alex_joni> emc1 that is..
[16:14:36] <jepler> alex_joni: yeah, that seems like a pretty strange thing to do to me
[16:14:38] <fenn> i thought you could run it in simulation mode
[16:14:52] <alex_joni> fenn: why would you want to do that?
[16:15:15] <fenn> er, um, nevermind
[16:15:32] <alex_joni> certainly not if you want to try the TP fix..
[16:15:40] <alex_joni> you won't see sh*t in simulation :))
[16:17:43] <Jacky^> alex_joni: do you think emc2 could be used to control a small robot ?
[16:17:50] <alex_joni> surely
[16:18:01] <fenn> as long as there are no if statements
[16:19:21] <Jacky^> alex_joni: could be nice if emc can run on a board like this: http://www.acmesystems.it/?id=705
[16:21:35] <alex_joni> if it's i386 no problem
[16:21:54] <Jacky^> :)
[16:22:09] <alex_joni> not sure about it if it's ARM based
[16:22:25] <Jacky^> i know ..
[16:22:39] <Jacky^> ive an ipaq with arm cpu :/
[16:29:38] <jepler> how much do those boards cost? They don't seem to have any american resellers.
[16:34:27] <Jacky^> Jacky^ is now known as Jacky^afk
[16:35:09] <alex_joni> jepler: hi there
[16:35:14] <alex_joni> still around?
[16:35:21] <jepler> alex_joni: yeah, off and on
[16:35:34] <alex_joni> do you read mail on the emc-devel list?
[16:36:04] <jepler> sometimes
[16:36:26] <alex_joni> I wondered what your opinion is on the mail I wrote
[16:36:30] <jepler> a lot of it goes over my head
[16:36:46] <alex_joni> mostly concearning coding
[16:36:46] <jepler> which message in particular?
[16:37:01] <alex_joni> the one about HAL driver standardisation
[16:37:12] <jepler> unfortunately, I know nothing about emc
[16:37:14] <jepler> er, emc2
[16:37:19] <alex_joni> never mind that..
[16:38:03] <alex_joni> but you got coding skills :)
[16:39:00] <alex_joni> otoh, now that I look at it, you need to look at some hal drivers, to know what I'm talking about..
[16:39:16] <alex_joni> so never mind.. I can understand if you don't have the time / nerve to do that
[16:39:20] <jepler> what do you mean when you say
[16:39:21] <jepler> HAL exported functions:
[16:39:21] <jepler> void board_name.<board_id>.dac_write
[16:39:30] <jepler> that doesn't make sense as a C or C++ declaration
[16:39:35] <alex_joni> right :)
[16:39:43] <alex_joni> it's a HAL function
[16:39:55] <alex_joni> inside HAL you got threads, and you can add functions to those threads
[16:40:06] <alex_joni> so that they get executed once / thread
[16:40:26] <alex_joni> * alex_joni goes home..
[16:40:31] <alex_joni> back online in 20 mins
[16:40:42] <alex_joni> maybe you'll be still around
[16:40:48] <jepler> I think the idea of providing abstract interfaces is a great one, but I don't know for sure if your ideas cover everything
[16:41:03] <jepler> I don't quite undertsand the inverted inputs either. why not just a 'bool is_inverted' instead of the union you have?
[16:41:26] <jepler> er, inverted digital I/O, in either direction
[16:41:30] <alex_joni> that struct gets kmalloc'ed
[16:41:39] <alex_joni> yeah.. that might be an idea
[16:41:49] <alex_joni> but it's easier to do it like this (for the user)
[16:42:00] <alex_joni> just connect the normal or the inverted signal
[16:42:15] <jepler> gotta go for now .. ask me again later and I might have more to say
[16:42:27] <alex_joni> if not he would need to set up a constant (either 5V or 0V) and connect that to the parm
[16:43:08] <Jymmm> Mornin Folks!
[16:43:09] <alex_joni> jepler: will do
[16:43:13] <alex_joni> bye Jymmm
[16:43:17] <fenn> heh
[16:47:16] <Jymmm> Who was that masked man?
[17:00:07] <Jacky^afk> Jacky^afk is now known as Jacky^
[17:13:46] <etla> hi
[17:19:08] <etla> anyone here ?
[17:21:07] <ValarQ> no
[17:21:36] <ValarQ> we are all out fishin ;)
[17:22:13] <etla> how do I zero an axis in axis ? :)
[17:23:21] <etla> axis wakes up with Y=2540 and will not move because "move out of range"
[17:23:30] <cradek> hit home
[17:23:45] <cradek> or maybe you have a g54 offset programmed?
[17:24:28] <etla> haven't got any home switches...
[17:24:43] <etla> zero offsets resets Y=2540
[17:24:53] <etla> (=100inches)
[17:26:08] <etla> show relative position = 100
[17:26:14] <etla> show machine position = 1.153
[17:26:17] <etla> whats that ??
[17:26:37] <cradek> uhh
[17:26:45] <cradek> I don't know...
[17:26:59] <cradek> on my emc1 system I built with NO_SWITCHES which makes HOME just zero the axis (causes no motion)
[17:27:12] <cradek> I have no idea how you get that behavior on paul's build
[17:28:45] <etla> in general, why would emc wake up with other than Y=0 ?
[17:28:46] <cradek> try in MDI mode: G10 L2 P1 X0Y0Z0
[17:29:27] <etla> that resets to X=0 Y=100 Z=0
[17:29:48] <etla> I don't have this behavior with tkemc or mini
[17:29:48] <cradek> now G0 X0Y0Z0
[17:30:03] <etla> linear move out of range
[17:30:19] <cradek> do you have a G92 offset or something?
[17:31:32] <etla> in manual mode 'zero offset' changes the X and Z coordinates by about -4
[17:32:16] <cradek> zero offset only affects the selected axis
[17:32:40] <etla> g92x0y0z0 did it ! what's g92 used for ?
[17:32:51] <cradek> for screwing stuff up
[17:33:01] <cradek> (seriously)
[17:33:07] <etla> ahaa... why would axis wake up with g92x100
[17:33:13] <etla> I mean y100
[17:33:26] <cradek> you had it there from a previous session
[17:33:45] <cradek> axis never uses g92 (of course you can type it at MDI)
[17:33:55] <etla> opening a file gives me wild g92 values again !
[17:34:13] <cradek> maybe you should exit and check your var file
[17:35:04] <cradek> 5220 should be 1, everything else should be 0
[17:35:10] <etla> emc.var is mostly zeros
[17:35:23] <etla> yes, 5220=1
[17:35:23] <cradek> is your system bdi?
[17:35:29] <etla> yes, 4.30
[17:35:49] <cradek> booting my bdi
[17:36:20] <cradek> my axis is slightly updated but I don't think it's anything relevant
[17:37:10] <cradek> is there anything special about your setup? is it metric?
[17:38:26] <etla> yes, inch ini file with RS274NGC_STARTUP_CODE = G21
[17:38:31] <etla> freqmod
[17:39:29] <etla> I seem to get different (random) g92 offsets when I restart emc
[17:39:48] <cradek> even before you load a file?
[17:40:17] <etla> yes.
[17:41:34] <cradek> I can't explain that
[17:41:40] <cradek> what version is your axis?
[17:41:44] <cradek> help/about
[17:42:19] <etla> 1.1rc1-20050925
[17:43:15] <cradek> we had some metric fixes on 20051011
[17:43:26] <cradek> maybe you should update
[17:43:35] <cradek> probably this does not affect your g92 problem though
[17:44:18] <xet7> What is newest BDI? 4.30?
[17:44:25] <cradek> yes
[17:44:31] <cradek> it's probably the last one for a while
[17:44:40] <xet7> why?
[17:44:46] <cradek> it finally works right
[17:44:52] <xet7> cool :)
[17:46:00] <cradek> etla: you can install the udpates to axis by just copying some files around - you don't have to install devel stuff to build it.
[17:46:36] <etla> ok, it's in some interpreted language and not compiled ?
[17:46:41] <cradek> yes
[17:47:18] <etla> reopen current file also resets my g92's to wild values
[17:48:36] <rcsu> the hal_introduction is a great document, but in the pdf version the images should be a bit larger :)
[17:48:46] <cradek> etla: with the new AXIS you can have your ini in terms of cm or dm or m, so you can have pretty much whatever scale you want and it will still work right.
[17:49:27] <cradek> etla: of course we haven't tested that much :-)
[17:50:34] <etla> OK... is that emc2 stuff or only axis related ?
[17:50:48] <cradek> it's only axis
[17:54:16] <cradek> etla: I put that RS274NGC_STARTUP_CODE in my ini, but it doesn't seem to do anything
[17:54:35] <cradek> does it for you?
[17:55:13] <etla> mdi etc wakes up in G21 mode
[17:55:27] <cradek> huh, mine doesn't
[17:55:28] <etla> the backplot display units need to be changed from the menu
[17:56:29] <etla> what's the blue sphere in the backplot window ?
[17:57:00] <cradek> the cyan thing? That's the absolute machine origin
[17:57:05] <cradek> it means you have an offset in effect
[17:57:24] <cradek> the relative origin is the three axes
[17:59:05] <etla> hmm... presser zero offset and axis seems to parse the loaded g-code file again. is that right ?
[17:59:11] <etla> pressed
[17:59:13] <cradek> yes
[17:59:22] <cradek> because it will need to move
[18:02:03] <etla> after a g92x0y0z0 , loading a files resets the g92 offsets to strange values
[18:04:19] <cradek> are you **sure** you examined the right var file
[18:04:20] <cradek> ?
[18:04:42] <cradek> on my bdi it looks like generic.var is the one that is used
[18:05:04] <cradek> I have to go for a while
[18:05:50] <etla> yep ! there are strange numbers in generic.var
[18:05:59] <etla> but hwo come these are not used with tkemc or mini ?
[18:26:33] <cradek> I don't know
[18:29:23] <etla> do you know how to set up limit switches as home switches ?
[18:29:34] <etla> i.e. need to move off them after tripping
[18:29:46] <cradek> nope
[18:29:51] <cradek> I don't know anything about switches
[18:29:59] <etla> ok
[18:30:01] <cradek> I thought I read somewhere that it can't be done, you need separate switches for home
[18:30:21] <etla> I tried resetting all values in generic.var but after restarting emc the strange values are there again !
[18:47:30] <alex_joni> greetings
[18:47:50] <Jymmm> bye alex_joni
[18:47:56] <alex_joni> hey Jymmm
[18:47:57] <alex_joni> :P
[18:48:16] <alex_joni> * alex_joni feels great :)
[18:48:23] <Jymmm> got laid?
[18:48:40] <alex_joni> Jymmm: that's a regular thing .. :)
[18:48:51] <alex_joni> got out of the shower, preparing for a massage
[18:48:58] <Jymmm> alex_joni: Yeha, but did this time involve someone else?
[19:15:38] <anonimasu> hello
[19:15:48] <Jymmm> olleh
[19:16:21] <anonimasu> what's up?
[19:16:29] <Jymmm> skipped steps
[19:18:38] <anonimasu> :/
[19:19:23] <Jymmm> anonimasu http://groups.yahoo.com/group/Xylotex/message/4083
[19:20:37] <anonimasu> so have you tried tweaking down your accels?
[19:22:04] <anonimasu> does the motor get warm?
[19:23:06] <anonimasu> brb going to go to the kiosk
[19:24:15] <Jymmm> anonimasu: red the foloowup at the bootom that link (I hate repeating myself =)
[19:24:18] <Jymmm> read
[19:24:46] <anonimasu> mm
[19:25:26] <Jymmm> fsck... I can't type today
[19:59:24] <les> hi ray
[19:59:32] <les> late lunch brealk for me
[19:59:40] <les> break
[20:01:53] <Jacky^> Jacky^ is now known as Jacky^afk
[20:28:29] <Jymmm> woof!
[20:33:10] <Jacky^afk> Jacky^afk is now known as Jacky^
[20:39:07] <alex_joni> 'lo
[20:39:50] <alex_joni> jepler: still around?
[20:40:42] <rayh> Hi alex_joni
[20:40:48] <alex_joni> hey rayh
[20:40:50] <rayh> Hi les
[20:41:03] <alex_joni> how's the mazak going?
[20:41:31] <rayh> Good. We have some motion. Still modifying the quill and spindle for encoder.
[20:41:42] <alex_joni> motion is good
[20:41:50] <alex_joni> controlled one is even better :D
[20:42:12] <rayh> John is working on the connection between hal and emc right now.
[20:43:22] <alex_joni> * alex_joni lends him some wires
[20:43:31] <alex_joni> oh.. and duct tape
[20:43:37] <alex_joni> LOTS of duct tape
[20:46:02] <alex_joni> * alex_joni tries to figure out a way to do the index-pulse homing
[20:46:11] <alex_joni> _any_ thoughts are welcomed
[20:48:30] <rcsu> how short is the pulse ?
[20:48:47] <alex_joni> 1 impulse / revolution
[20:48:59] <alex_joni> so a typical encoder (about 2000 impulses / revolution).. you do the math
[20:49:06] <alex_joni> 360 / 2000 = degrees
[20:49:14] <alex_joni> pretty short.. but that's not the problem
[20:49:24] <alex_joni> most of the boards have support for it
[20:49:33] <alex_joni> but there are 2 ways to do it:
[20:49:50] <alex_joni> 1. when the index pulse is received the counter gets reset
[20:49:59] <rcsu> but you dont know the actual position on startup
[20:50:02] <alex_joni> 2. when the indec pulse is received the counter gets saved
[20:50:17] <rcsu> so homing can only be done with a switch
[20:50:20] <alex_joni> these two are done in hardware, given you're in a certain homing sequence
[20:50:30] <alex_joni> rcsu: that part is pretty much sorted out
[20:50:35] <alex_joni> look at the HAL docs
[20:50:40] <alex_joni> * alex_joni looks for a link
[20:50:57] <rcsu> i read the hal_intor
[20:51:02] <rcsu> intro
[20:51:31] <rcsu> i didnt get your prob
[20:51:34] <alex_joni> EMC2_Code_Notes.pdf
[20:52:00] <alex_joni> http://www.linuxcnc.org/EMC2_Code_Notes.pdf
[20:52:01] <alex_joni> page 21
[20:52:05] <alex_joni> can you look?
[20:52:22] <rcsu> sure
[20:53:07] <alex_joni> ok.. last 2 cases
[20:53:26] <alex_joni> the whole homing is done like there
[20:53:36] <alex_joni> what is not clear is what happens when the index pulse comes...
[20:53:58] <alex_joni> I start from this premise:
[20:54:25] <alex_joni> you cannot expect to see the index pulse in software, and get a chance to remember the current position
[20:54:31] <alex_joni> without getting a sligth error
[20:54:40] <alex_joni> and do that on all 6 axes at once
[20:55:31] <alex_joni> and this is a known fact.. so hardware manufactureres (like the STG, or any other servo controlling card) have taken care of this in hardware
[20:58:17] <alex_joni> rcsu: still there?
[20:59:15] <rcsu> yep, thinking about that
[20:59:26] <alex_joni> till now it's ok
[20:59:36] <alex_joni> we didn't get to the part where the problem is :)
[20:59:56] <rcsu> k, go ahead
[21:00:09] <alex_joni> like I said .. need to do it in hardware
[21:00:24] <alex_joni> but in that case one of 2 things happen
[21:00:36] <alex_joni> either case 1. the counter gets reset on the index pulse
[21:01:03] <alex_joni> or case 2. the counter gets saved on the index pulse, and you can use it later on
[21:01:18] <alex_joni> now.. for me working on a driver either one is very easy to do
[21:01:25] <alex_joni> the problem is connecting the result to emc
[21:01:38] <alex_joni> or better said making emc aware of those
[21:02:40] <alex_joni> so.. case 1.
[21:02:53] <rcsu> hm, i would use case 2
[21:03:04] <alex_joni> emc will get the correct position after the index, but PID will totally freak out
[21:03:18] <rcsu> you have a constant in-(de-)creasing counter
[21:03:21] <alex_joni> imagine PID tracking the position, and the counter suddenly goes from 25321 to 0
[21:03:22] <alex_joni> :D
[21:03:49] <rcsu> 25321+1
[21:03:50] <alex_joni> case 2: pro- no problem with PID
[21:04:04] <alex_joni> con- need a lot of hooks to connect the hal driver to emc
[21:04:21] <alex_joni> I need current position, saved position, do_home, homed
[21:05:00] <alex_joni> emc drives do_home, the driver looks at index pulse, when it is tripped it reports the saved position and homed
[21:05:45] <rcsu> hmm, to complicated (at the first look)
[21:05:52] <tcpsyn> Hello
[21:05:58] <alex_joni> * alex_joni is open to suggestions
[21:06:03] <alex_joni> tcpsyn: SYN_SENT
[21:06:11] <alex_joni> err.. hello
[21:06:12] <alex_joni> :)
[21:06:13] <tcpsyn> heh.
[21:06:22] <tcpsyn> Guess this isn't for EMC hardware.
[21:06:27] <tcpsyn> I think I got the wrong room
[21:06:29] <tcpsyn> doh.
[21:06:31] <alex_joni> not really
[21:06:32] <alex_joni> :)
[21:06:49] <tcpsyn> ah well. Thanks for being friendly.
[21:06:59] <alex_joni> huh... ?
[21:07:45] <tcpsyn> oh
[21:08:05] <tcpsyn> alright then.. check it out.
[21:08:08] <alex_joni> as in EMC=Enhanced Machine Controller
[21:08:17] <tcpsyn> damnit
[21:08:25] <alex_joni> just to make it sure that you don't mean another EMC :)
[21:08:31] <tcpsyn> I'm talking in EMC=High availabilty storage systems
[21:08:35] <tcpsyn> bah
[21:12:03] <alex_joni> rcsu: back to where we were
[21:13:28] <rcsu> alex_joni: am i right when you can capture the index pulse but not the exact absolute position at the mmt when the pulse comes ?
[21:13:46] <alex_joni> you cannot capture the index pulse
[21:13:55] <alex_joni> it might just go by, because of speed
[21:14:07] <alex_joni> but usually there is a latch on the board that tells you
[21:14:12] <rcsu> ok, so not latched
[21:14:19] <rcsu> grr
[21:14:30] <rcsu> thats what i mean
[21:14:32] <alex_joni> rcsu: yes
[21:14:33] <alex_joni> :)
[21:15:06] <anonimasu> hello
[21:15:17] <alex_joni> hey anders
[21:15:22] <alex_joni> you come just right
[21:15:24] <alex_joni> read back :)
[21:16:23] <rcsu> reading the average error on the pulse distance is the first thing i would do
[21:16:48] <alex_joni> what do you mean?
[21:17:24] <rcsu> say, you see the first index @ 950
[21:17:36] <rcsu> the second is @ 1025
[21:17:39] <anonimasu> hm
[21:17:44] <anonimasu> alex_joni: what part of it?
[21:17:55] <rcsu> the third is @ 980
[21:17:56] <alex_joni> anonimasu: how would you do it
[21:18:05] <alex_joni> rcsu:that's not a bad idea :)
[21:18:24] <rcsu> the result is somewhat under 1000
[21:18:26] <alex_joni> but not sure how it'll help
[21:18:35] <alex_joni> ahh.. like that? .. nah
[21:18:50] <alex_joni> I thought to check for the index pulse (get it latched at 937)
[21:19:01] <alex_joni> then check for the next one (get it latched at 2937)
[21:19:07] <anonimasu> *tries to comperhend to do what*
[21:19:19] <alex_joni> anonimasu: the problem is with index pulses
[21:19:31] <anonimasu> I would go by the index pulse
[21:19:31] <alex_joni> how would you connect the indexpulse to emc
[21:19:39] <alex_joni> to emc I said
[21:19:45] <anonimasu> yeha
[21:19:48] <anonimasu> err
[21:19:54] <alex_joni> to do the driver is easy
[21:20:03] <alex_joni> and check for the index pulse, and report it
[21:20:08] <alex_joni> the problem is what happens then
[21:20:24] <anonimasu> yeah
[21:20:49] <rcsu> hm why should emc know that ?
[21:20:54] <anonimasu> are you talking about the procedure or the actual connection into it..
[21:21:00] <alex_joni> both
[21:21:09] <alex_joni> the homing sequence is documented
[21:21:40] <alex_joni> so that might count as procedure, although it doesn't say what happens to the counters, and I'm trying to figure that out now
[21:21:53] <alex_joni> rcsu: the whole point is to find the HOME position
[21:22:01] <alex_joni> so emc needs to know it has reached it :)
[21:22:07] <alex_joni> or at least where it is
[21:22:55] <rcsu> an index pulse doesnt tell anything about a home position
[21:23:22] <rcsu> i would keep it in the hal driver, no where else
[21:23:45] <alex_joni> I agree.. but it is used in the homing sequence
[21:23:49] <alex_joni> so how do you report it?
[21:23:58] <alex_joni> only as a bit: index has been tripped?
[21:24:04] <alex_joni> won't give you any precision
[21:24:06] <rcsu> and the homing sequence is done in emc ?
[21:24:11] <alex_joni> YES
[21:24:12] <alex_joni> :P
[21:24:26] <alex_joni> in the motion controller
[21:24:40] <alex_joni> HAL is a Hardware Abstraction Library
[21:24:50] <rcsu> yep, i know
[21:24:53] <alex_joni> it only connects things to other things
[21:25:00] <alex_joni> doesn't do anything more than that
[21:26:13] <rcsu> thats what the words say 'hardware abstaction', and imho HOME is a really hardware bound thing
[21:26:27] <alex_joni> nope it's not
[21:26:32] <Jacky^> alex_joni: where is your video section in the web ?
[21:26:37] <alex_joni> it's a concept, a position of the machine
[21:26:42] <alex_joni> Jacky^: what video?
[21:26:47] <Jacky^> i'm looking for a video with steppers and wires ..
[21:26:56] <Jacky^> ive seen time ago ..
[21:27:32] <alex_joni> http://dsplabs.utt.ro/~juve/emc/mytoy/
[21:27:42] <Jacky^> yeah ! thanks :)
[21:30:01] <alex_joni> anonimasu: got an idea?
[21:32:01] <anonimasu> the homing sequence I have a good idea about
[21:32:13] <alex_joni> yes, but that's pretty well done
[21:32:31] <alex_joni> the problem is how you figure out when the index pulse has been tripped
[21:32:45] <alex_joni> the easiest way is to have it reset the counter
[21:33:02] <alex_joni> but that will totally bust the PID
[21:33:04] <anonimasu> hm, isnt storing the position into memory easier..
[21:33:25] <alex_joni> both are easy from the driver's point of view
[21:33:32] <alex_joni> but you still need to pass it to emc
[21:33:52] <anonimasu> and then when you leave the pulse in the oposite direction | goes low then you zero the position..
[21:34:33] <alex_joni> you can't presume you can see the index as it is
[21:34:41] <alex_joni> you'll see a latched version
[21:34:54] <alex_joni> for example: on the STG you can only watch 2 index pulses at once
[21:34:57] <anonimasu> yeah, but you could have the driver trigger on change..
[21:35:01] <alex_joni> out of the 8 that there are
[21:35:18] <alex_joni> so you might be 4 periods away from the actual position
[21:36:06] <anonimasu> yeah but that's a hardware problem
[21:36:21] <anonimasu> I would go over the index pulse then sneak back slowly.. waiting for it to fall..
[21:36:55] <alex_joni> hardware has support for precise indexing
[21:37:03] <alex_joni> either counter reset, or counter latching
[21:37:19] <alex_joni> so when the index is tripped the current counter is saved to another location
[21:37:25] <alex_joni> which your driver can read
[21:37:42] <alex_joni> the problem is .. what do you do with it?
[21:38:11] <alex_joni> you report it to emc as it is? (home-pos)
[21:38:13] <alex_joni> not nice, but simple
[21:38:18] <anonimasu> hm, why would you do anything you are zeroed..
[21:38:21] <anonimasu> eg homed..
[21:38:33] <alex_joni> you are not
[21:38:41] <alex_joni> you're missing the step I'm talking about
[21:38:55] <alex_joni> let's say we are in the homing sequence.. very advanced
[21:39:03] <alex_joni> driving slowly towards the index pulse
[21:39:27] <alex_joni> (bear in mind this can be on 6 axes simultaneously)
[21:39:36] <alex_joni> ok.. now the index is hit, the counter gets saved in the STG hardware
[21:40:07] <alex_joni> on the next period the driver will see that the index is tripped, and report an index-tripped through HAL back to emc
[21:40:27] <alex_joni> but untill that happens, the axis might have moved 12 counts
[21:41:17] <anonimasu> how fast are you moving?
[21:41:25] <anonimasu> that's what it depends on..
[21:41:33] <alex_joni> don't know.. might be 500 counts / sec
[21:41:38] <alex_joni> or 50000 counts /sec
[21:42:10] <alex_joni> still it takes up to 4 period's for the stg to even know the index pulse has been tripped
[21:42:18] <alex_joni> then another one to report it to emc
[21:42:32] <anonimasu> yeah but cant you stop moving after it's latched..
[21:43:02] <alex_joni> nope
[21:43:11] <alex_joni> that needs to come from emc too
[21:43:25] <alex_joni> and that also takes some time to propagate, so you can't stop exactly there
[21:43:43] <alex_joni> but remember you have the exact home position stored
[21:44:15] <anonimasu> wouldnt moving back there work?
[21:44:39] <alex_joni> back where?
[21:44:45] <anonimasu> it's to the exact home position..
[21:45:07] <anonimasu> sorry I am slow today..
[21:45:08] <anonimasu> :/
[21:45:11] <alex_joni> well.. if you have the exact home position in emc, you don't care about anything else
[21:45:26] <alex_joni> you know where it is, substract it from the current, and you're done
[21:45:30] <anonimasu> yeah
[21:45:45] <alex_joni> the problem is how to get it to emc
[21:45:47] <anonimasu> just offset the current position from the home position..
[21:45:50] <anonimasu> yeah,
[21:46:50] <anonimasu> but how is that done today?
[21:46:58] <alex_joni> on emc1 ?
[21:47:01] <anonimasu> emc2
[21:47:06] <alex_joni> not done
[21:47:07] <alex_joni> :D
[21:47:08] <anonimasu> ah
[21:47:22] <alex_joni> that's why I'm frying my head ..
[21:47:30] <alex_joni> I could just hack it up in 10 mins and have it working..
[21:47:36] <anonimasu> :/
[21:47:39] <alex_joni> but I'd rather make it the best possible way
[21:47:43] <anonimasu> yeah
[21:48:29] <anonimasu> you can read the hal stuff from within anywhere almost right?
[21:48:40] <alex_joni> probably
[21:48:48] <alex_joni> but certainly from the motion controller
[21:49:04] <alex_joni> but you need to export further pins (for the position), and that is ugly :)
[21:49:57] <anonimasu> yep
[21:50:42] <alex_joni> huh.. I just stumbled upon some crazy stuff
[21:50:46] <alex_joni> :-O
[21:51:07] <anonimasu> :D
[21:51:08] <anonimasu> tell me about it
[21:51:18] <alex_joni> activeX talking to emc
[21:51:27] <alex_joni> * alex_joni steps aside
[21:51:31] <anonimasu> lol
[21:51:43] <alex_joni> not to be striken down by the gods
[21:52:15] <alex_joni> there was thunder nearby
[21:52:16] <alex_joni> :D
[21:54:16] <anonimasu> *yawns*
[21:54:33] <alex_joni> I think that's the only way..
[21:54:42] <alex_joni> have a bidirectional signal
[21:55:03] <alex_joni> emc sets it for homing-indexing operation
[21:55:28] <icee> so, perhaps it was performance related. I replaced the machine that I saw lots of EMC crashes on and I've only had one thus far.
[21:55:31] <alex_joni> the board pulls it down when index has been tripped, and reports the saved home position
[21:55:33] <icee> on the new hardware.
[21:55:45] <alex_joni> icee: I see :)
[21:55:55] <icee> And I think that one was to some shutdown weirdness the time before.
[21:56:06] <icee> I do have another problem tho now
[21:56:23] <icee> I can't turn down the acceleration parameter too -low- or i get following errors.. and i'm not sure why that would be
[21:57:38] <alex_joni> running steppers?
[21:57:51] <alex_joni> also: running emc1 ?
[21:57:54] <alex_joni> or BDI ?
[21:58:25] <rcsu> alex_joni: btw, i got emc2 to run on my machine
[21:58:38] <alex_joni> rcsu: nice, so it was that lapic stuff..?
[21:58:43] <rcsu> alex_joni: next step is to have the nc-step card going
[21:58:49] <icee> yah, running steppers, on emc2
[21:59:04] <rcsu> alex_joni: you put it into the manu.lst :)
[21:59:04] <alex_joni> rcsu: so you're in the same boat as Imperator_? ?
[21:59:09] <icee> the emc2 branch checked out last night, though I observed this on the other machine too, from emc2 from a couple weeks ago
[21:59:13] <alex_joni> rcsu: yes
[21:59:28] <alex_joni> icee: hmm..
[21:59:43] <alex_joni> following errors mean the pulse gen. can't keep up
[21:59:56] <alex_joni> but low accel shouldn't do that..
[21:59:56] <icee> the pulse gen can easily keep up with much higher pulse rates
[22:00:06] <alex_joni> what means low?
[22:00:30] <icee> i have my max speed set to 0.25 now, and i want to run a max acceleration of 1 or so
[22:00:38] <icee> under 4 doesn't work
[22:00:44] <icee> it seems to 'overshoot on deceleration', maybe
[22:00:57] <icee> accelerating up isn't a problem, but when it goes to turn a corner, etc..
[22:02:11] <alex_joni> that's inches I presume?
[22:02:18] <icee> INPUT_SCALE is 32000, so my peak pulse rate is 8000/second
[22:02:18] <icee> yes
[22:03:55] <alex_joni> ok.. jsut checking to be sure, not that you work in attometers
[22:03:57] <alex_joni> :D
[22:04:01] <icee> ;)
[22:04:10] <alex_joni> or exa :)
[22:04:14] <icee> I can provide the ini if you want to see if you can reproduce it
[22:04:23] <alex_joni> ok.. mail it over
[22:04:27] <alex_joni> alex.joni AT robcon.ro
[22:04:31] <icee> one moment
[22:06:09] <icee> OK, that reliably breaks very quickly on cds.ngc
[22:07:48] <cncuser> hidiho
[22:08:39] <cncuser> hi alex_joni :)
[22:08:45] <alex_joni> icee: An attachment named emc.ini was removed from this document as it
[22:08:45] <alex_joni> constituted a security hazard. If you require this document, please contact
[22:08:45] <alex_joni> the sender and arrange an alternate means of receiving it.
[22:08:50] <alex_joni> cncuser: hello
[22:08:52] <icee> haha
[22:09:00] <icee> should I just include it in the text of the email?
[22:09:06] <alex_joni> yup
[22:09:08] <cncuser> renate mit too txt
[22:09:08] <cradek> ha, I've seen that before, sounds like mimedefang
[22:09:15] <cncuser> rename it to txt
[22:09:20] <alex_joni> let's arrange alternate means of receiving it.
[22:09:23] <alex_joni> cradek: yup
[22:09:32] <cncuser> ini is some windows stuff, maybe thats why its consider dangerous
[22:09:37] <alex_joni> oh.. chris :)
[22:09:50] <cradek> I protect dozens of windows machines with that software
[22:10:00] <alex_joni> lurking for long? or you just turned up?
[22:10:01] <icee> when using it by hand, in axis.. i can hold down the arrow and it's fine
[22:10:06] <icee> when i let it go it says following error
[22:10:11] <cncuser> cradek: does it let txt attachments throu ?
[22:10:25] <alex_joni> cncuser: it's configurable
[22:10:29] <cradek> cncuser: probably, but it's configurable
[22:10:44] <cradek> mine is hacked to do all sorts of clever things like look inside zips
[22:10:54] <cncuser> then hopefully there was noone so stupid to configure such thing :)
[22:11:26] <cradek> icee: that's not axis-specific
[22:11:34] <cncuser> hi cradek, btw :)
[22:11:38] <cradek> hello
[22:11:39] <icee> cradek: I know. Just being clear.
[22:11:49] <icee> I'm sure in tkemc or whatever it'd blow up too.
[22:12:03] <cradek> yeah. is this steppers?
[22:12:08] <jepler> cradek: it's possible that it is axis specific in this respect: we used to not follow the ini for jog speed, while the other FEs did...
[22:12:24] <cradek> oh, that's right
[22:12:35] <cradek> I always assume that as soon as I fix something in cvs, everyone has it
[22:12:53] <alex_joni> but the error happens on running programs too
[22:12:58] <alex_joni> so that's not axis-dependent
[22:13:02] <icee> well, it doesn't just happen for jogging, yah
[22:13:24] <cradek> icee: something about your config is definitely marginal
[22:13:36] <cradek> icee: accel, period, pid, ...?
[22:13:48] <icee> cradek: I was saying I can't set acceleration as low as i'd like.
[22:13:53] <alex_joni> cradek: I'll forward you the ini
[22:13:53] <icee> maybe a trajectory planner bug?
[22:14:14] <icee> alex: can you reproduce the problem?
[22:14:23] <cradek> icee: there seem to be all sorts of interdependent things and emc only works well when they're in certain ranges
[22:14:24] <alex_joni> icee: first I'd reduce the TRAJ accels too
[22:14:32] <cncuser> cradek: i am doing a developer image of the iso. what python version should i use ? is 2.4.2 ok, or are you bound to 2.3.x ?
[22:14:32] <cradek> icee: none of us really understand it, I think.
[22:14:45] <icee> alex: okay, let me try that
[22:14:51] <cradek> I believe anything 2.2 or newer is fine
[22:14:56] <cncuser> cool
[22:15:10] <cradek> what do you mean by developer image of the iso?
[22:15:21] <alex_joni> you have 20 for accel in TRAJ (default and max), and 0.3 in axis_0
[22:15:37] <alex_joni> cradek: I think one that you can do programming on
[22:15:39] <cradek> .3??
[22:15:50] <icee> yah, don't the axis ones override the traj ones?
[22:16:12] <cncuser> cradek: its based on puppy 1.0.5 with development environment. i want to add everything to make it possible to get latest cvs stuff, compile it, and output an iso.
[22:16:19] <icee> ah, that fixes it. Guess I misunderstood.
[22:16:25] <cncuser> emc2, rtai, axis
[22:16:28] <alex_joni> icee: glad to help :P
[22:16:34] <cradek> cncuser: cool
[22:16:35] <icee> cradek: 0.3 is degenerate.. i really want 1.0 or so
[22:16:44] <icee> thanks people :)
[22:16:53] <alex_joni> icee: I don't have much knowledge about traj & co.. but as I know it:
[22:16:56] <cncuser> but will be around 500mb i assume :(
[22:17:09] <alex_joni> the value in TRAJ needs to be max(axis_0,1, etc)
[22:17:16] <cradek> cncuser: I really think the user should have a built anonymous checkout. When he wants to update, he only has to type "cvs up" and rebuild.
[22:17:22] <alex_joni> seems stupid to have it, instead of figuring that out..
[22:17:34] <alex_joni> cradek: great idea
[22:17:41] <alex_joni> easy to explain too
[22:17:43] <cncuser> cradek: yes, i thought of something like that too
[22:17:51] <cradek> IMO, there's no reason for the thing that comes on the CD to be different from a cvs checkout.
[22:17:55] <icee> alex: well, in this case, min(axis_0, ...)
[22:18:05] <alex_joni> not min, max
[22:18:23] <icee> well, if it's greater than the axis one, in this case
[22:18:24] <icee> it breaks
[22:18:24] <cradek> with BDI it's a nightmare to rebuild. Half the time, the thing that you're running isn't even IN cvs.
[22:18:24] <cncuser> cradek: the proiblem iss, the size of a development environment exceeds the minimal livecdkonzept
[22:18:25] <alex_joni> all your axes are at 0.3
[22:18:33] <alex_joni> so traj should be 0.3
[22:18:46] <icee> alex: I have a feeling if i set axis2 to max accel 20
[22:18:48] <icee> and traj to 20
[22:18:49] <cradek> cncuser: yeah, I can sure see that might be a problem
[22:18:52] <icee> it'll still break on axis 0 and 1
[22:18:52] <alex_joni> but if you'd have one axis at 0.3 and one at 10, then maybe 10 should be the value
[22:19:01] <cncuser> cradek: so i think one userversion (precompiled, around 20mb iso download) and a development version is the way i go
[22:19:03] <alex_joni> icee: try it out :P
[22:19:06] <icee> OK, will confirm.
[22:19:14] <alex_joni> icee: cool
[22:19:24] <alex_joni> * alex_joni still wants to hijack the conversation
[22:19:27] <cradek> cncuser: I bet most people are not scared off by a 100MB or more download
[22:19:29] <cncuser> updates can be done easaly via squasfs images and the unionfs feature
[22:19:34] <alex_joni> back to encoders & index pulses
[22:19:47] <alex_joni> cradek: not for a puppy
[22:19:53] <alex_joni> that's a way too fat puppy
[22:20:02] <alex_joni> that puppy needs exercise :))
[22:20:03] <cncuser> cradek: cool thing is... you can run it enterely from ram, that means, more resistance against hardwarefailure
[22:20:05] <icee> yup.. it breaks
[22:20:16] <alex_joni> icee: try 0.3 then :P
[22:20:28] <cradek> my whole issue is that I would like for any user to be able to get a bugfix I make in cvs.
[22:20:35] <alex_joni> but I bet the axis with 10 will never accel like that
[22:20:43] <alex_joni> cradek: right
[22:20:49] <alex_joni> so CVS for both?
[22:20:49] <icee> yah, i'll just set them all the same
[22:20:53] <cncuser> cradek: the lets make a buildscript that compiles it and make binaries available ?
[22:20:54] <alex_joni> emc & axis ?
[22:21:04] <cncuser> those can be easily upgraded
[22:21:14] <alex_joni> cncuser: I can host(cron) that
[22:21:32] <cradek> axis is easy to upgrade, usually even without building
[22:21:41] <alex_joni> so always latest emc2 & axis inside the iso
[22:21:45] <alex_joni> this sounds fun
[22:21:46] <cncuser> cradek: for the developerversion i think the cvs update is perfect
[22:22:09] <icee> isn't there a benefit to having .. nice stable releases with known sets of bugs and many people testing them?
[22:22:11] <cncuser> cradek: ok
[22:22:16] <cncuser> cradek: what changes ?
[22:22:24] <icee> so someone doesn't get the ISO when a stupid bug is committed and have problems and come here begging for help
[22:22:24] <alex_joni> icee: no, we want bleeding edge
[22:22:34] <cradek> cncuser: usually only the scripts (not the C++ module)
[22:22:35] <alex_joni> latest kernel, latest rtai, buggy software
[22:22:43] <cncuser> cradek: sorry, i just wake up. *laugh* its python ;)
[22:22:43] <alex_joni> don't sleep well at night :D
[22:22:44] <cncuser> brrr
[22:22:46] <icee> oh wait.
[22:22:50] <icee> got another following error
[22:22:51] <cncuser> haha
[22:22:52] <icee> it just took longer
[22:22:53] <icee> agh
[22:23:07] <alex_joni> hssss
[22:23:48] <alex_joni> icee: no other idea :/
[22:23:48] <cncuser> cradek: then i only need cvs. or why not put it on a http accessible cvs. then i can use wget
[22:24:18] <icee> I'll keep tuning it. It's closer to working than ever before.
[22:24:19] <cncuser> hmm, i start a wiki on that right now :)
[22:24:40] <jepler> I don't really want to offer anon cvs of axis, I don't think I know enough to operate it safely
[22:24:57] <cradek> you can already use webcvs to get certain files
[22:24:57] <alex_joni> jepler: but you offer tar.gz
[22:24:58] <alex_joni> so that's ok
[22:25:03] <cradek> sure, or that
[22:25:16] <alex_joni> * alex_joni wants to set up the same for emc2
[22:25:26] <alex_joni> what the heck.. I'm doing it right now
[22:25:36] <cncuser> a unpacked daily tarball with correct timestamps is easyli to sync via wget
[22:25:48] <cncuser> :)
[22:26:00] <alex_joni> logger_aj, buggeroff
[22:26:00] <alex_joni> I'm logging. I don't understand 'buggeroff', alex_joni. Try /msg logger_aj help
[22:26:26] <alex_joni> logger_aj: so why don't you respond to ssh you crappy 'puter?
[22:26:26] <alex_joni> I'm logging. Sorry, searching removed.
[22:28:12] <rcsu> /etc/init.d/cyrus stop
[22:28:38] <alex_joni> cyrus service : stopped
[22:29:55] <cncuser> i love unionfs
[22:30:03] <rcsu> eehhh, wrong channel, sorry
[22:30:14] <alex_joni> rcsu: :P
[22:31:12] <cncuser> if i know it earlier id have hundreds of filesystem patches made up for easy install and removal :)
[22:31:24] <cncuser> laying around :)
[22:31:50] <alex_joni> lol
[22:35:30] <Jacky^> hey guys
[22:35:50] <Jacky^> do you think 5 phase stepper are necessary for this machine ?
[22:35:55] <Jacky^> http://www.carlsondesign.com/
[22:35:57] <cncuser> well, gotta go, traminer wine party :)
[22:36:02] <cncuser> cu later
[22:36:06] <CIA-6> 03jmkasunich * 10emc2/configs/mazak_rf.hal: more work on the mazak
[22:36:06] <alex_joni> later cncuser
[22:36:07] <Jacky^> or another motor type can be used
[22:36:55] <alex_joni> 6-phase AC servo
[22:37:08] <alex_joni> 5300W minimum
[22:37:45] <icee> I'm designing the motor driver for those 1kW servos
[22:37:50] <icee> i'm mildly scared of it.
[22:38:07] <icee> I think I'll be able to build a driver for $150/axis or so
[22:38:43] <alex_joni> icee: hope those aren't the ones doing 0.3 i/sec^2 ?
[22:38:44] <alex_joni> :D
[22:38:57] <icee> nope, nope. hope to do 10G's :)
[22:39:01] <cradek> icee: there are only 53 of them left
[22:39:06] <cradek> icee: I hope you bought yours already
[22:39:17] <icee> yup, i did. allegedly they've shipped
[22:39:42] <icee> wow.. they've sure sold them quick.
[22:43:49] <icee> I bought one of the 300W servos, too. I'm thinking about putting it on the spindle on my mini mill.
[22:44:14] <les> hi guys
[22:44:21] <Jymmm> * Jymmm is just having a rough week....
[22:44:50] <jepler> $150 per axis for the driver? That's no small change.
[22:44:56] <Jymmm> one of my clamps came loose and the piece moved... I *think* I was able to recover though.... we'll see.
[22:45:02] <icee> jepler: it is much, much cheaper than buying commercial ones.
[22:45:13] <jepler> yeah, but so much more expensive than steppers
[22:45:23] <Jymmm> hi les
[22:45:29] <icee> jepler: the 3 phase bridge for the IGBT alone is $70
[22:45:39] <icee> jepler: yah, but these are 1 kilowatt servos.
[22:45:47] <alex_joni> jepler: not sure you have 1kW steppers
[22:45:52] <alex_joni> at least useable :)
[22:46:08] <jepler> alex_joni: no, of course not.
[22:46:08] <icee> the three phase bridge IGBT that is
[22:47:44] <icee> jepler: $150 for the driver + $40 for the kilowatt servo per axis compares very favorably to stepper solutions
[22:47:48] <alex_joni> so.. if you go with servos it's expensive
[22:48:05] <alex_joni> usually a 1kW axis is at about 2-3k USD
[22:49:03] <icee> I plan to use the pico systems interface for my board
[22:49:47] <les> I guess that's why my 105 servos went so fast
[22:49:59] <les> about 1kw
[22:50:24] <les> $300-500 per axis depending on the sero amp
[22:50:29] <les> servo
[22:50:59] <icee> i expect this to be about $200/axis total-- amplifier, motion control board. (well, not including PC)
[22:51:12] <les> well that would be great
[22:51:24] <icee> not counting opportunity cost of building/testing all this stuff myself
[22:51:27] <les> to bad there isn't more motors
[22:51:28] <icee> and hopefully not killing myself in the process
[22:51:51] <icee> 340VDC == teh scary
[22:51:59] <les> the opportunity cost is gazillions of course
[22:52:35] <alex_joni> icee: we switched to 600V servos on our robots lately
[22:53:07] <icee> alex: eek. what are the voltages inside the amplifiers?
[22:54:04] <icee> you must need like 2kVDC at substantial current to drive that well
[22:55:28] <alex_joni> the power to the servo is 600V DC
[22:55:38] <alex_joni> perhaps I wasn't clear.. :)
[22:55:51] <alex_joni> it chops it up, and generates AC from that
[22:55:55] <icee> oh, OK
[22:56:00] <alex_joni> frequency controlled
[22:56:28] <icee> I'm just using 340VDC because it's probably big enough and easy to get (just rectify + filter 220VAC)
[23:04:28] <alex_joni> icee: nice
[23:08:24] <les> hmm saw a post from Jon where he claims 10K+ servo updates per second with emc
[23:08:27] <les> ????
[23:08:47] <Jacky^> anonimasu: you around ?
[23:08:50] <alex_joni> loong seconds :)
[23:08:54] <les> heh
[23:16:03] <alex_joni> cradek: you there?
[23:19:06] <alex_joni> hmm.. did you guys know about : http://cvs.sourceforge.net/cvstarballs/emc-cvsroot.tar.bz2 ?
[23:19:12] <alex_joni> it gets updated nightly
[23:22:47] <anonimasu> yes
[23:22:53] <anonimasu> * anonimasu is getting mad at the little avr
[23:23:01] <anonimasu> bit_is_set() seems to be broken
[23:23:09] <alex_joni> huh
[23:23:16] <alex_joni> maybe it's bent?
[23:23:36] <anonimasu> if ( !bit_is_set(PORTA,0) ) {//do stuff}
[23:23:39] <anonimasu> wont work..
[23:23:43] <anonimasu> seems bent
[23:24:13] <alex_joni> try if (bit_isnt_set(PORTA,0) ) {//do stuff}
[23:24:17] <alex_joni> lol
[23:24:31] <anonimasu> heh
[23:25:03] <anonimasu> and how would that help me
[23:27:34] <Jymmm> anonimasu : try if (bit_is_bent(PORTA,0) ) {//do stuff}
[23:27:40] <anonimasu> heh
[23:35:40] <alex_joni> back in 10 mins
[23:37:15] <les> oh my. the computer models of wilma are diverging horriibly
[23:37:27] <les> My folks are up from florida
[23:37:53] <Jymmm> Hi Mom!
[23:37:59] <Jymmm> and Pops!
[23:38:02] <les> trying to decide whether to rush back.
[23:38:32] <Jymmm> did they just get in today?
[23:38:49] <cradek> is there time to get in and back out?
[23:39:11] <les> From what I see It will be another one of those Houston things....with hundreds of thousands of out of gas cars on the interstate
[23:39:21] <cradek> yep
[23:39:22] <les> they have been here for several days
[23:40:07] <les> need contigiency planning so one lane is left over for contiuous roving fuel trucks
[23:41:03] <les> bet they won't do it
[23:43:13] <icee> les: do they at least reverse the 'other side' of the highway to be leaving also?
[23:43:41] <les> who knows if they will do contraflow
[23:43:48] <les> just talked to folks
[23:43:53] <les> they will stay
[23:44:07] <les> going back is kinda the wrong way
[23:44:24] <les> but other relatives may need to come up
[23:44:50] <Jymmm> PARTY AT LES'
[23:44:50] <les> the computer models basically collapsed in the last run
[23:45:52] <les> the gfdl model showed a 1650 nautical mile change in the 5 day position over the last run 6 hrs ago
[23:46:06] <les> in other words
[23:46:09] <les> who knows
[23:47:05] <icee> always when things tangentially interact with a landmassk it makes things a lot less certain
[23:47:25] <les> fluid flow: the last great unsolved problem in physics
[23:47:39] <icee> i think by tomorrow morning you'll know fairly definitively where it's going
[23:47:40] <Jymmm> les : Just add LOX
[23:47:46] <les> haha
[23:47:49] <icee> oh, fluid flow is understood, it's just chaotic
[23:47:56] <les> heh
[23:47:57] <icee> you can never measure to enough precision to know exactly what's going to happen
[23:48:44] <les> yeah we call that the "butterfly flapping it's wings in china" problem
[23:49:04] <les> otherwise known as initial condition sensitivity
[23:49:04] <icee> i think i heard about that in jurassic park.
[23:49:28] <Jymmm> Butteryfly Effect was a movie out a couple years ago
[23:49:35] <les> hmmm
[23:50:52] <les> oh and as far as party at les'.....
[23:51:03] <les> party on I-75 is more like it
[23:51:03] <icee> les: from the noaa human-prepared 5 day forecast cone, you guys look relatively safe
[23:51:19] <Jymmm> lol
[23:52:08] <les> for south central florida... I see 25% prob of hurricane winds...10% prob of devastating wind
[23:52:12] <les> good odds?
[23:52:26] <icee> les: aren't you in northern georgia?
[23:52:45] <les> yes but My folks and kids live in south florida
[23:52:51] <les> oh we are fine here
[23:52:58] <icee> oh, i see :/
[23:53:17] <icee> well, by tomorrow morning it ought to be a lot more clear.
[23:53:22] <les> hope so
[23:53:55] <les> The interstate parking lot is the big issue...
[23:54:47] <les> I could run down to board up...not sure.
[23:55:47] <cradek> to me, it seems too late for that
[23:56:22] <LawrenceG> icee, hello... I would like to toss over driver ideas for the surplus servos
[23:56:27] <cradek> it would be hell to get stuck on the highway like that
[23:56:35] <cradek> with all those desperate people
[23:56:39] <icee> lg: I've been studying that exact problem for hte past couple days :)
[23:56:53] <les> cradek: perhaps. The model timing is so uncertain...
[23:57:24] <LawrenceG> icee, ... I bought 3 of the 300watt units (no breaks) and hope to replace the size 34 steppers on my machine
[23:57:32] <LawrenceG> brakes
[23:58:08] <cradek> I also would love to replace my NEMA23 steppers
[23:58:09] <icee> lg: well, I'm designing a massive DC power supply (that you probably would want to size down) and looking at how to build the output bridges out of IGBTs
[23:58:24] <cradek> I hope you guys come up with something (I don't know squat about servos)
[23:58:30] <LawrenceG> icee, I have been poring over 3ph bldc driver app notes for the last few days
[23:59:05] <icee> lg: it really is pretty simple. figure out phase angle by encoder pulses
[23:59:30] <icee> run closed-loop control on current via pwm output duty cycle at that angle
[23:59:34] <LawrenceG> icee, motorola and microchip have quite a few apps notes on bldc drives, but not actually in torque servo applications... more like vfd's
[23:59:52] <icee> lg: let me find you the parts i'm looking at