#emc | Logs for 2007-12-28

Back
[00:00:22] <BigJohnT> yes
[00:00:43] <fenn> that will work just fine (if the vfd accepts single phase)
[00:02:23] <Gamma-X> ok good
[00:02:54] <Gamma-X> i jsut wanna make sure the vfd will not act as a static phase converter buy limiting me to 2/3 the power of the motor
[00:03:03] <fenn> no, it won't
[00:03:12] <BigJohnT> a vfd is not a static phase converter
[00:03:25] <BigJohnT> a static phase converter is just a bunch of caps
[00:03:36] <fenn> just a bunch of craps
[00:03:42] <BigJohnT> that
[00:03:47] <BigJohnT> s what i meant
[00:04:21] <BigJohnT> a 3hp vfd will run a 3 hp motor with a little to spare
[00:04:52] <BigJohnT> I doubt you will load the 3hp motor to it's max ever unless you screw up your g code
[00:05:16] <BigJohnT> and drive the tool holder into the work... how do I know that... hmmmm
[00:07:32] <BigJohnT> 79% gamma
[00:09:21] <gezr> 5, 18, 15, hmm, I believe your going to want to find what is supplying the thing in page 24, its the power supply for the control
[00:10:08] <gezr> ah ha
[00:10:13] <gezr> on page 25 is the key
[00:10:52] <gezr> on the very left of the page, you will notice, there are connections for either 110 or 220v operation
[00:11:54] <gezr> yeah, thats a transformer
[00:12:11] <fenn> sho 'nuf
[00:12:34] <fenn> too bad it doesnt say wire colors
[00:13:26] <gezr> colors dont matter :)
[00:13:50] <fenn> oh right, its just terminals on the transformer
[00:14:04] <Gamma-X> look at u guys go!
[00:14:11] <Gamma-X> thank god for people like u guys.
[00:14:18] <fenn> * fenn yawns
[00:14:22] <Gamma-X> i realy appreciate it lol im serious
[00:14:22] <gezr> I used to do this stuff for a living
[00:14:31] <Gamma-X> thanks in advance guys.
[00:14:37] <gezr> but since I wasnt making any money at it im not doing it anymore
[00:14:45] <Gamma-X> true
[00:15:09] <BigJohnT> before I can even get it downloaded you guy have it figured out
[00:15:19] <Gamma-X> lol BigJohnT get faster internet
[00:15:32] <Gamma-X> and dont live in teh sticks!
[00:15:57] <Gamma-X> brb
[00:16:22] <BigJohnT> love it in the sticks I can pee off my deck and no one will see
[00:16:43] <gezr> 17 is also a power supply, for +5 and +/-18v
[00:16:54] <fenn> this is the transformer, i assume the black/white wires are attached to a circuit board with the terminal numbers labeled http://fennetic.net/pub/irc/gamma-x/DC061008003.jpg
[00:17:03] <JymmmEMC> http://youtube.com/BigJohnT_pissing_off_the_deck/
[00:17:24] <jmkasunich> google satellite saw him!
[00:17:46] <BigJohnT> need the hubble to see that LOL
[00:17:54] <JymmmEMC> rotf
[00:18:12] <BigJohnT> ROFLMAO
[00:18:36] <BigJohnT> too many trees here jymmm
[00:19:20] <gezr> fenn yes upper right
[00:19:29] <fenn> actually black/white is probably something else, and what we want it out of view
[00:19:31] <gezr> on the far left, is another power supply
[00:19:47] <fenn> yep that's the 5/-18
[00:20:13] <gezr> fenn is this your machine or gamma's?
[00:20:36] <fenn> gamma's
[00:20:52] <gezr> the schematics however do not show, servo drives, nor spindle drives, so his vfd/converter search may still be on
[00:22:17] <fenn> gezr: the idea is to run the spindle from a vfd and rewire the servos to run on single phase of some sort
[00:22:29] <gezr> 90v 10amp servo drives probably
[00:22:42] <laguille> ok i have the complete errors (first when emc starts second with the dmesg cmd) where can i post it?
[00:22:54] <fenn> gezr: 140V .. 30 amp peak?
[00:23:05] <BigJohnT> on my anilam I just plugged the whole thing into a wall outlet as the tech at anilam told me was best to do...
[00:23:32] <ALS> laguille:pastebin.org
[00:24:08] <BigJohnT> pcb 494 drives the servos
[00:24:17] <laguille> ok the i do that... thanks
[00:25:33] <gezr> bigjohn and that fine for the servos ? they dont need 3 phase?
[00:25:45] <BigJohnT> nope
[00:25:54] <BigJohnT> just the spindle
[00:26:26] <fenn> i only see encoder read stuff, no power electronics
[00:26:28] <gezr> so im going to go out on a limb here, and bet his spindle has some sort of existing manual speed contrl on it?
[00:26:39] <BigJohnT> yea
[00:26:50] <gezr> oh hell this is a no brainer then
[00:26:54] <BigJohnT> yea
[00:27:16] <BigJohnT> no rigid tapping going on here
[00:27:25] <BigJohnT> or there as it would be
[00:27:34] <ALS> longilander
[00:27:38] <gezr> right, just a simple 110 control, with a spindle on and off
[00:27:45] <BigJohnT> yep
[00:27:53] <gezr> probably a contactor, so yeah, all he needs is a vfd
[00:28:14] <BigJohnT> the only thing I did was add a potential relay to drop out the servos if I loose 3phase to my spindle
[00:28:18] <laguille> oops i have post the two errors without any other comments from line 1 to 292 error when emc2 starts and from ln 300 to the end of the post the dmesg command? thanks in advance
[00:28:22] <BigJohnT> kinda ugly if you don't
[00:28:31] <gezr> I perfer mitsubishi, because I know how to program those
[00:29:07] <gezr> 5kw good enough for 3hp right?
[00:29:29] <gezr> maybe Ive got that backwards, 1.5kw =3hp?
[00:29:46] <BigJohnT> hmmm, don't have a chart here, my other brother john could tell you off the top of his head
[00:30:09] <laguille> http://pastebin.org/13091
[00:30:30] <gezr> well, he is gonna need a bit larger cause he is feeding single, but not much
[00:31:07] <laguille> kk
[00:31:56] <BigJohnT> generally the biggest single phase to 3 phase is 3hp
[00:32:45] <fenn> laguille: what does 'lsmod | grep parport' say?
[00:33:12] <BigJohnT> i like grep now...
[00:33:43] <fenn> BigJohnT: you'll like 'locate' too
[00:33:48] <gezr> hmm
[00:34:08] <gezr> fenn looks like its in the modules?
[00:34:32] <gezr> wait your right, paraport is in use by something else
[00:35:09] <BigJohnT> hmmm, locate what does it do
[00:35:29] <fenn> its like grep, but it looks for file names anywhere on your disk
[00:35:33] <BigJohnT> opening a terminal window as we type
[00:35:38] <laguille> ?? my english is not perfect?
[00:35:52] <gezr> updatedb, then try which :)
[00:35:59] <BigJohnT> mine is not either
[00:36:04] <fenn> laguille: open a terminal and type lsmod | grep parport
[00:36:41] <BigJohnT> the gs2 from automationdirect 3hp is 2.2kw
[00:37:32] <BigJohnT> locate is cool too
[00:37:43] <BigJohnT> I like linux now
[00:38:06] <laguille> en français its return "command introuvable" unknow command
[00:38:47] <fenn> perhaps you typed an I instead of L?
[00:39:02] <laguille> yes i type i .. wait
[00:39:33] <fenn> the second | is shift-\
[00:39:35] <BigJohnT> intruvable = impossible to find
[00:39:37] <gezr> fen look at lines 215-225 or so
[00:40:04] <gezr> kernel is auto loading the port
[00:40:22] <fenn> gezr: yes unfortunately i'm not smart enough to know exactly what is going on from that
[00:40:41] <laguille> parport_pc 30916 1 and parport 29640 3 ppdev,lp, parport_pc
[00:40:47] <fenn> yay
[00:40:53] <gezr> 182.021583] ppdev: user-space parallel port driver
[00:41:40] <fenn> laguille: those modules should not be loaded, and i dont know why they are
[00:41:54] <gezr> config para port not as a module im betting
[00:42:04] <gezr> laguille did you compile this kernel?
[00:42:38] <fenn> laguille: a temporary solution is to do: rmmod lp ppdev parport parport_pc
[00:43:09] <fenn> or whatever order in which they need to be removed
[00:43:16] <laguille> i think i compile this kernel 6 month ago... butafter upgrading i dident reboot te pc ?
[00:43:33] <fenn> ah
[00:43:36] <fenn> just reboot then
[00:43:54] <laguille> ok i reboot fist...
[00:43:57] <gezr> hold on
[00:44:11] <gezr> lets check his .config really quick
[00:44:20] <fenn> eh, who cares about his config
[00:44:21] <BigJohnT> dinner bell has rung here in swamp east missouri
[00:44:46] <gezr> laguille : could you type cd /usr/src/linux
[00:44:50] <fenn> gezr: when you install the emc package it downloads a realtime kernel package and sets it to boot as default
[00:44:59] <gezr> oh
[00:45:29] <fenn> so, unless he needs a custom kernel i'm not going to worry about it
[00:45:37] <fenn> and then i'm still not going to worry about it :)
[00:45:57] <gezr> heh
[00:46:32] <gezr> I think he may have compiled the kernel on his own, and left the paraport set to kernel and not module, but we shall see
[00:46:54] <fenn> no, it wouldnt show up in lsmod then
[00:46:57] <gezr> fenn : so you dont do kernel compiles?
[00:47:07] <fenn> no
[00:47:19] <fenn> i have, but it is no fun
[00:47:29] <gezr> there is a kernel directive for auto loading modules, kauto loader or something
[00:48:00] <gezr> its no big deal
[00:48:10] <gezr> and your right
[00:48:38] <fenn> i used to have to compile kernels before i switched from redhat to debian
[00:48:45] <fenn> but redhat sucks
[00:48:57] <fenn> and the ubuntu stuff works on debian just fine
[00:49:03] <laguille> it's run !!!!!![dtc]
[00:49:19] <gezr> yes, 4.0.3-1ubuntu5)) #1 Fri Jun 9 20:51:19 EEST 2006 hehehe, june
[00:49:24] <laguille> after rboot emc 2.2.2 run???
[00:49:38] <gezr> execlent
[00:49:51] <gezr> fenn you were right, I didnt notice the date earlier
[00:50:40] <laguille> thank you very much fen and gezr and every body:-p
[00:51:09] <laguille> sorry for the wrong smile... you suggest me to test something?
[00:52:20] <ALS> laguille: what type of machine? you run
[00:52:43] <laguille> a mill retrofiting
[00:52:52] <ALS> knee?
[00:54:17] <Gamma-X> i think im gunna buy this hitachi if i can do a vfd, http://cgi.ebay.com/3-HP-VFD-VARIABLE-FREQUENCY-AC-MOTOR-SPEED-DRIVE_W0QQitemZ140192704353QQihZ004QQcategoryZ71393QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[00:59:10] <Gamma-X> any thoughts on that for my machine?
[00:59:43] <gezr> it doesnt quite match what you need for a spindle
[01:00:11] <fenn> it doesnt?
[01:01:05] <gezr> it says its not recomended for low speed aplications
[01:01:09] <fenn> oh pfff
[01:01:41] <laguille> with emc2.0.5 i had problem to configure the ini file according to my mecanic, the linear unit parameter was a numer not metric or inch ?? and the input scale was unknow for me...??? i could run the machine in manual but in automatic l alway the error message: machine exess the minimum/or maximun limit on each axis... now i hope i can configure the machine with step conf... thank again for your Great Help...
[01:01:56] <gezr> read this thread, gamma-x http://www.cnczone.com/forums/showthread.php?t=42068
[01:04:42] <laguille> ALS: what a knee?
[01:09:10] <ALS> knee mill bridgeport and the like's
[01:11:34] <ALS> laguille: what part of the rock you from?
[01:12:19] <Gamma-X> fenn u guys come up with a soulution for me? lol
[01:12:39] <fenn> Gamma-X: yea take some wires off, plug them back in
[01:13:07] <gezr> gamma-x : thats pretty much what its going to take
[01:13:14] <Gamma-X> lol
[01:13:27] <Gamma-X> uhhh wich wires
[01:13:27] <Gamma-X> ?
[01:13:34] <fenn> the ones going into the transformer
[01:13:35] <gezr> you have an electrician friend?
[01:13:38] <Gamma-X> yup
[01:13:44] <gezr> get him to come over
[01:13:54] <Gamma-X> the transformer inside the servo box?
[01:14:02] <Gamma-X> or power panel
[01:14:14] <fenn> i thought there was only one
[01:14:41] <Gamma-X> the big black thing in uppper right hand corner of servo box?
[01:14:54] <gezr> gamma-x basically you just need to get 110v to the control, and make sure that its not set to use 220v
[01:15:14] <Gamma-X> ok
[01:15:24] <gezr> and make sure your not feeding power back to the spindle from that stuff either
[01:15:26] <fenn> Gamma-X: you also need to disconnect it from the 3-phase lines
[01:16:12] <Gamma-X> ok
[01:16:17] <gezr> basically, that transformer on the right is getting 2 lines input, I do not know if they are 220v, or 110v because im not there, but your electrician friend will be able to tell
[01:16:38] <Gamma-X> ok
[01:16:46] <Gamma-X> i know what ur talkin about i got up and close last night
[01:16:55] <fenn> gezr: if you run a motor at 2x speed but in a lower gear, it shouldnt bog down when you go back to "normal" motor speed
[01:16:56] <laguille> ALS: yes that's a old knee mill, i'm from Switzerland and you?
[01:17:09] <gezr> if IM looking at it correctly, the right side of the transformer will have its wires removed, and replaced with a line from your power socket, either 110v single, or 220v single, thats the only remaining question
[01:17:38] <ALS> laguille: eastern USA
[01:17:52] <gezr> gamma-x you can test this, do yu have a voltage meter?
[01:18:04] <Gamma-X> pretty sure yeah
[01:18:24] <gezr> crap, re post the photo please
[01:18:25] <Gamma-X> power to servos is 140v so im assuming its only 120 ac
[01:18:44] <gezr> what makes you say power to servos is 140v?
[01:18:59] <fenn> i bet it wants 240 because the 3-phase is 240, so it would have to step down to get to 110AC/140DC
[01:19:18] <fenn> but if you can wire it up either way, doesnt really matter
[01:19:20] <gezr> yeah, thats what im thinking
[01:19:22] <Gamma-X> http://imagebin.org/12618
[01:19:46] <Gamma-X> gezr it says it on the servos lol
[01:20:08] <Gamma-X> ill try it with 120 and if it dont work i add 220..
[01:20:11] <Gamma-X> no harm right?
[01:20:11] <gezr> no,
[01:20:14] <gezr> do not
[01:20:20] <Gamma-X> lol
[01:20:56] <fenn> disconnect the 3-phase wires first, and test it with the multimeter to make sure it's disconnected
[01:21:12] <gezr> it will smoke the transoformer if its set at 220v, the circuits behind it will draw a ton of current to make up for the lack of voltage, and unless you have a device to catch the smoke, youll never get it back in there and once its gone its gone
[01:21:35] <gezr> gamma-x, what I would do, is label all the wires on that transformer and whre they go, then remove all of them
[01:21:50] <Gamma-X> they already are
[01:21:57] <gezr> then put 110v on the left side, where the original wire were, and measure the voltage on the left side
[01:22:09] <Gamma-X> ok
[01:22:19] <fenn> which left? :)
[01:22:23] <gezr> just take those wires off, so your just dealing with the transformer
[01:22:49] <Gamma-X> if im correct, on the right side of the transformer there is 1 wire that is bridged to another terminal
[01:22:53] <gezr> the wires on the top left of the transformer, appear to be ground wires
[01:23:03] <gezr> then you need to give it 220v
[01:23:03] <Gamma-X> i think x2 is connected and then bridged to x1
[01:23:34] <gezr> if it has an pattern like this |__| thats 220v
[01:23:46] <Gamma-X> yup
[01:23:47] <gezr> if it looks like this |X| then its 110
[01:23:51] <Gamma-X> well
[01:23:53] <Gamma-X> idk
[01:24:10] <Gamma-X> i dont know what u mean by those wonderful illustrations lol
[01:24:23] <Gamma-X> http://www.driveswarehouse.com/Drives/AC+Drives/Sensorless+Vector+VFD/SV022iC5-1.html
[01:24:26] <Gamma-X> that or the hitachi
[01:24:33] <gezr> okay, on the transformers right side connector plate, there are 4 connector points right?
[01:24:38] <fenn> |U| 220V
[01:24:39] <gezr> 4 screw terminals that is
[01:25:32] <gezr> if we call them 1 2 3 4, then 220, is going to have wire--1 wire-on-2-to-3 wire --4
[01:26:00] <gezr> wait a moment
[01:26:04] <gezr> thats not right
[01:26:11] <gezr> im thinking 440
[01:26:40] <Gamma-X> yeah 4 connectors
[01:26:41] <fenn> it's just series or parallel
[01:27:04] <fenn> if its series the ratio is 2:1, if its parallel the ratio is 1:1
[01:27:41] <Gamma-X> 1 wire on terminal 2 and a single wire connecting terminal 2 to termnial 1
[01:27:46] <fenn> once you have the transformer isolated from the circuit you can measure the voltage on the output
[01:28:04] <gezr> yeah, just do that gamma, remove the wires, but not the jumper, and test
[01:28:20] <laguille> more thanks for your support, emc.2.2.2 seem to run properly, tomorrow (here its 2 oclock on the morning) i will connect the pc to the machine and hope can configure emc2 for my cnc...:D see you next time tanks tanks tanks
[01:28:33] <gezr> laguille good luck
[01:28:50] <Gamma-X> the schematics didnt say? lol
[01:29:21] <gezr> not really
[01:29:38] <fenn> the schematics say 'hook it up this way for 220 and that way for 110'
[01:30:16] <gezr> fenn, on lower voltate stuff the jumpers can look like an L then backwards L instead of the cris cross pattern
[01:31:45] <fenn> meh, draw it with loopie loops for all i care
[01:31:58] <Gamma-X> lol
[01:32:02] <fenn> its the same circuit
[01:32:44] <fenn> gezr: oh are you talking about bus bars?
[01:33:07] <gezr> yeah
[01:33:49] <ALS> *ALS steps out his door and looks south east for the light show
[01:33:51] <gezr> the jumpers on the transformer, instead of an X or a __ it can be a |_ _|
[01:34:14] <fenn> ALS: rocket launch?
[01:34:33] <ALS> Gamma_X
[01:34:38] <Gamma-X> holla at ur kitten
[01:34:52] <Gamma-X> im just kidin whats up
[01:35:10] <ejholmgren> my kitten got old and started fighting back
[01:35:13] <Gamma-X> lol
[01:36:55] <Gamma-X> ALS whatrs up
[01:37:09] <Gamma-X> ohh ur waiting for me to blow up
[01:37:21] <Gamma-X> i live in ny so look far
[01:38:24] <ALS> gamma_X: you'll get it don't hurry make sure your safe
[01:38:39] <Gamma-X> lol
[01:38:50] <fenn> looks like nasa's got the same strategy on sts-122
[01:38:51] <ALS> I'm syracuse not to far
[01:40:29] <fenn> * fenn shuts up about space shuttles
[01:40:49] <ALS> good old birds
[01:41:04] <fenn> well, they're old all right
[01:41:10] <gezr> what is a space shuttle?
[01:41:40] <ALS> money pit
[01:41:55] <fenn> gezr: it's what you get when you dont know whether to make a reusable launcher, a cargo rocket, or an airplane
[01:42:04] <gezr> * gezr is confused nasa is movie production company, and shuttle is just a show they produce right?
[01:42:32] <gezr> by the way, I would love to work for NASA
[01:42:54] <fenn> i would love to get money to build spaceships
[01:42:55] <ejholmgren> I'm a tax paying american ... they should put that b*tch up every 6 months
[01:43:12] <Gamma-X> ejholmgren then ur taxes wqould double
[01:43:28] <fenn> bah
[01:43:49] <ejholmgren> I live in MN ... then my federal would finally approach my state taxes
[01:43:49] <Gamma-X> fenn so i can deffanatly use a vfd though?
[01:44:03] <gezr> gamma-x only for your spindle
[01:44:09] <Gamma-X> http://www.driveswarehouse.com/compare.aspx?pid=226&pid=1214
[01:44:18] <Gamma-X> gezr but i dont need 3 phase for anything else right?
[01:44:24] <Gamma-X> wich one should I get..
[01:44:54] <gezr> right
[01:46:54] <fenn> they are both 'sensorless vector' so whats the difference
[01:47:01] <Gamma-X> features
[01:47:07] <ALS> Gamma_X: the pretty one
[01:47:08] <Gamma-X> the little stuff
[01:47:21] <Gamma-X> is it worth an extra 50 bucks for some little extra crap
[01:47:38] <Gamma-X> the white one puts out more power
[01:48:29] <fenn> it might be worth it, just for the external inputs for run/stop
[01:49:27] <fenn> heh i like this line: The braking torque at capacitive feedback is the average deceleration torque at the shortest deceleration
[01:49:43] <Gamma-X> lol
[01:49:59] <Gamma-X> fenn my machine already has that.
[01:50:01] <gezr> you try to stop one too fast youll learn quickly how to diganose one of those magic boxes
[01:50:18] <Gamma-X> gezr my machine also has a power brake
[01:50:30] <gezr> then make sure your vfd can talk with it
[01:50:51] <Gamma-X> gezr hmm
[01:51:15] <Gamma-X> gezr theres a solenoid but its controlled by the controller lol
[01:51:27] <gezr> thats exciting news
[01:51:33] <Gamma-X> hahahah
[01:51:42] <Gamma-X> i meant like it doesnt need to talk to it.
[01:51:53] <Gamma-X> if the power is off on the motor then its off and the brake comes on.
[01:52:41] <ALS> electro magnetic brake?
[01:52:43] <gezr> well, your right, so your gonna have to find a way then to get what the controler does to talk with the vfd, and stick a pot on the side of the control box to control speed, but the 350 model has more options for your set up
[01:52:55] <Gamma-X> ALS no a phnuematic brake
[01:53:09] <ALS> ah
[01:53:25] <Gamma-X> gezr its an old controll lol idk how i could do it like that
[01:53:47] <Gamma-X> gezr AT FIRST IM JUST USING THE VFD AS A 3 phase converter.
[01:53:47] <gezr> gamma-x basically what im talking about is this, you remove the input/output from the existing motor contactors in terms of the power lines
[01:53:50] <Gamma-X> nutin else.
[01:54:14] <Gamma-X> ok
[01:54:26] <gezr> then you have the signal that makes the vfd run forward, run to the top of the forward contactor, and vise versa for the reverse, then down from the og contactor to the vfd
[01:55:23] <fenn> the one on the right is prettier
[01:55:50] <Gamma-X> fenn but 50 bucks more lol
[01:56:13] <gezr> vfd gets it 220v sinle in, 3 wires out to the motor, whatever the output is fromt he vfd probably +10v to the top of each contactor, then back to the respective fwd, rev points on the vfd, then you run the speed wires from vfd to a potentiometer and make sure its outputing 0-10v when you twist on it, and your done
[01:56:44] <gezr> thats assuming the control just yanks down on a contactor to make the motor whirl
[01:56:54] <gezr> in its current state
[01:57:19] <gezr> oh yeah, your gonna have to program that puppy too
[01:57:25] <gezr> the vfd that is
[01:57:27] <Gamma-X> lol
[01:57:28] <fenn> oo scary
[01:57:37] <gezr> easy mode
[01:57:52] <gezr> just put your brian cap on and read the book lots
[01:58:38] <gezr> its giong to have all metric inputs and such so watch that, and if you need a braking resistor, I recomend finding some old stoves, and using those pan coils to do the work
[01:58:44] <Gamma-X> gezr but all i realy need to do is ruin the wires from my amchine to the vfd and not worry about the vfd reversing because the controller will do that.
[01:59:12] <gezr> no the controler wont do that, if the controler already does all that, then you have a vfd in there already
[01:59:27] <Gamma-X> gezr no.
[01:59:42] <gezr> if your machine is just pulling down contactors, then the contactors voltages are not going to make your vfd happy
[01:59:42] <Gamma-X> gezr on the machines is 4 buttons, reverse, stop, start, and forward.
[01:59:54] <ALS> magnetic contactor
[01:59:55] <Gamma-X> its jsut a switch from the 3 phase to the motor
[02:00:13] <gezr> yeah, which is what I described
[02:00:38] <Gamma-X> yeah thats there already but no vfd
[02:01:20] <gezr> your switch, selects wich contact is going to be pulled down, ie where the contactor voltage is going to go, its probably 110, so you need to keep that the same, and instead of having the contactor start the motor, the contactor meerly sends the go forward, or go reverse signal back to the vfd, from the vfd
[02:01:49] <SkinnypuppY34> Hi guys, I've got an Axis question, when touching off only the . ,the decimal key will work on the number pad with num lock on, only the top row keys over the letter keys will enter digits in touchoff box. Is is like this for everyone or a me issue?
[02:02:34] <gezr> that I dont know, im sorry
[02:02:37] <Gamma-X> gezr yeah but thats jsut more wires weverywhere
[02:03:16] <Gamma-X> im trying to stop a mess from happening and keep it simple
[02:03:35] <ALS> SkinnypuppY34: I think everyone I mostly use tkemc though
[02:04:27] <fenn> SkinnypuppY34: do you get a 'bad character' error when you try to use the number keys?
[02:04:41] <gezr> gamma-x, its 5, but yes, oh wait a moemnt, look, im just trying to keep you from pulling wires, since you have a switch and 2 push buttons, you can remove the existing wires from those, remove the contactor blocks, and put your vfd there, then use new wires to run the control lines, im just tying to keep you from sending a bad voltage into the vfd's control side
[02:05:17] <Gamma-X> ok
[02:05:51] <Gamma-X> gezr im just worry about the accesorys on the machine gettin power such as the coolant pump, oil pump, air brake.
[02:05:58] <Gamma-X> etc
[02:06:02] <fenn> it's captain harlock!
[02:06:03] <gezr> just look at how things are ran, if the control starts and stops the motor, then you want to learn how it does that, that way, you dont have to push buttons, you just need to control the speed
[02:06:07] <[Harlock]> hi all... got an issue with my parallel port
[02:06:11] <[Harlock]> yeah hehehe
[02:06:38] <gezr> gamma-x : are you following me?
[02:06:45] <[Harlock]> He was such a nice outlawed hero
[02:06:49] <Gamma-X> gezr to an extent yes
[02:07:21] <SkinnypuppY34> fenn it's stranger than I first thought that keypad will do 5 0 . - * / and + but not 1234678and9 ...
[02:07:37] <[Harlock]> I'm running a CVS snapshot from Oct 2007 of EMC2, and I can't figure why my parallel port isn't working
[02:07:56] <gezr> gamma-x : the vfd requires certian things to happen before it will start the motor, and those inputs must be, within the parameters of what the vfd wants
[02:08:07] <Gamma-X> damn lol
[02:08:19] <[Harlock]> I hooked up a 4 channels scope to it and I'm trying to get the HAL to toggle the pins and it does nuttin
[02:08:42] <gezr> gamma-x : this isnt a problem, its just a matter of how you want to do it, that vfd is perfect, the 350 buck one
[02:08:56] <fenn> SkinnypuppY34: i bet it is interpreting the keys as arrows or pageup/home/end etc
[02:08:56] <SkinnypuppY34> I don't get bad character but it will say bad number format if I give it some bogus syntax like ..05
[02:09:28] <gezr> gamma-x mind if I ddc chat with you for a bit?
[02:09:34] <Gamma-X> sure
[02:09:52] <fenn> SkinnypuppY34: oh, also, kde had some weird mode where you use the number pad to move windows around.. is that happening?
[02:10:36] <SkinnypuppY34> Yep in the MDI section 248and 6 move the highlighted section through mdi history and cursor left/right
[02:11:11] <fenn> [Harlock]: can you read a +5V signal on one of the input pins? (use a pullup resistor)
[02:12:02] <[Harlock]> fenn: Right now I have pins 2,3,4 and 5. they're all high
[02:13:13] <fenn> inputs dont need a pullup, its just so you dont damage the port if it's acting as an output
[02:14:07] <[Harlock]> I was planning on using pins 2 to 7 for the three axis
[02:14:20] <fenn> [Harlock]: is it an add-on port or just a regular 0x378 motherboard port?
[02:14:27] <[Harlock]> controlling steppers of course
[02:16:42] <[Harlock]> regular 0x378
[02:16:45] <fenn> um.. you aren't running with --enable-simulator are you?
[02:17:07] <[Harlock]> is it a compile time switch^
[02:17:08] <[Harlock]> ?
[02:17:11] <fenn> yes
[02:17:16] <[Harlock]> d'oh
[02:17:19] <[Harlock]> letme check
[02:17:32] <gezr> hey fenn, schematic link again?
[02:18:04] <fenn> gezr: http://fennetic.net/pub/irc/Crusader%20II%20Wiring%20Diagrams.PDF
[02:20:53] <fenn> that's neat that you can run the latency-test without a realtime kernel
[02:21:57] <[Harlock]> what do you mean by running "the latency-test without a realtime kernel"??
[02:22:20] <fenn> well, i'm here on just a regular computer, and i can run latency-test and it will tell me the latency
[02:22:40] <fenn> [Harlock]: you can tell if you did --enable-simulator by looking in src/config.log
[02:22:55] <fenn> but i think it would throw an error about unable to load parport module
[02:23:05] <fenn> so that's probably not it
[02:23:21] <[Harlock]> well I had to unload parport, ppdev, lp and parport_pc
[02:23:34] <fenn> hmm
[02:23:38] <[Harlock]> actually my kernel keeps reloading them automaticrappy at boot time
[02:23:44] <fenn> what kernel are you using?
[02:24:00] <[Harlock]> 2.6.20.2
[02:24:10] <[Harlock]> the only thing that worked for me up to now
[02:24:13] <fenn> did you compile it yourself with the adeos and rtai patches?
[02:24:52] <[Harlock]> on debian etch yes, following instructions on EMC2 etch compilation wiki
[02:25:05] <[Harlock]> I think Adeos is integrated with RTAI, isn't it?
[02:25:16] <[Harlock]> sorry I did this two months ago, and got carried away
[02:26:11] <fenn> adeos is.. the part of the kernel that talks to RTAI? or something
[02:26:21] <[Harlock]> oh... and I have the wonderful NVIDIA GeForce4 Ti4400... *sigh*
[02:26:46] <fenn> if it were me, i'd install the ubuntu packages (they work for me, on etch)
[02:26:57] <[Harlock]> works now, but I really had a hard time with 2.6.17
[02:27:14] <[Harlock]> you mean kernel and everything?
[02:27:31] <fenn> yes, there arent too many extra packages though
[02:28:00] <[Harlock]> I ran ./configure --enable-run-in-place --enable-build-documentation
[02:28:07] <Ziegler> hey gents... I am having problems finding the latest source... I complied the latest cvs ok, but it seems it does not want to run.
[02:28:10] <[Harlock]> so no simulation only flag
[02:28:37] <[Harlock]> Ziegler: I had something too from the cvs... running latency checks don't pass
[02:28:38] <cradek> Ziegler: the latest source is always the cvs trunk, and each release has a corresponding tag in cvs
[02:28:58] <Ziegler> err... latest stable source
[02:29:24] <Ziegler> im only looking to run sim
[02:29:55] <Ziegler> (you all have a good Christmas?)
[02:29:57] <[Harlock]> fenn: shouldn't I be able to simply toggle parport pins with halcmd?
[02:30:41] <fenn> Ziegler: hey where'd my wiki instructions for installing emc2-sim go
[02:30:53] <Ziegler> ??
[02:31:06] <Ziegler> does it say I edited it?
[02:31:50] <fenn> oops, sorry that wasn't supposed to be addressed directly to you
[02:31:53] <Ziegler> pheeww...
[02:31:59] <Ziegler> had me worried for second
[02:32:08] <fenn> [Harlock]: yes
[02:32:26] <Ziegler> emc/task/emctaskmain.cc 2612: can't initialize motion << this is what I get with the latest CVS when I attempt to run it
[02:32:29] <[Harlock]> I know the port is working. Linux toggles the pins on and off during boot time
[02:33:02] <fenn> grr stupid usemod deleted the old revisions too
[02:33:52] <fenn> [Harlock]: sure it's linux doing that? some bios'es test the port
[02:34:25] <fenn> [Harlock]: anyway i dont think i can help, i dont know that much about parports
[02:34:34] <[Harlock]> well i check the pin status during booting. The bios might be doing something, but Linux is definately booting at this stage
[02:35:20] <[Harlock]> I have to dig parameters on the hal_parport module
[02:35:39] <fenn> did you try running one of the sample configs?
[02:35:56] <[Harlock]> yes I started with stepper_mm
[02:36:03] <[Harlock]> from the emc2 interface
[02:36:57] <[Harlock]> but there's almost a dozen of parameters on parport.0 that seems to be timing stuff
[02:38:04] <[Harlock]> I might need probe_parport
[02:38:38] <fenn> Ziegler: anyway, edit your /etc/apt/sources.list to point to emc2-sim and get emc2.2-sim
[02:38:53] <fenn> or maybe the other way around
[02:39:55] <Ziegler> cd ..
[02:39:58] <Ziegler> oops
[02:40:06] <fenn> [Harlock]: do you have another computer to test on? perhaps you could just swap in the hard drive
[02:40:10] <[Harlock]> '/$
[02:40:18] <[Harlock]> nah...
[02:40:45] <[Harlock]> there's a nice info when rgrepping for "hal_parport" in probe_parport
[02:41:05] <Ziegler> fenn I dont have a apt/ dir under the emc install
[02:41:06] <[Harlock]> it says to try loading probe_parport first if hal_parport doesn't work right
[02:41:23] <[Harlock]> Ziegler: /etc/apt/
[02:41:26] <fenn> Ziegler: in the root of the filesystem
[02:41:40] <Ziegler> oops... read that as emc
[02:42:26] <fenn> deb http://www.linuxcnc.org/emc2/ dapper emc2.2-sim
[02:42:38] <fenn> sudo apt-get update; sudo apt-get install emc2-sim
[02:43:26] <Ziegler> (not on a system with apt installed)
[02:44:27] <Ziegler> all I am looking for is latest stable source.
[02:45:09] <fenn> cvs -z5 -d:ext:[email protected]:/cvs co -rv2_2_branch -demc2.2-branch emc2
[02:45:20] <Ziegler> danke fenn
[02:45:52] <maddash> suppose I set [traj]axes=4. is it possible to jog the 4th axis whilst the other three are in coord mode?
[02:46:29] <cradek> all joints switch from free to coordinated mode together
[02:46:41] <[Harlock]> loading probe_parport in halcmd sure toggles the port pins.. there's hope!
[02:48:00] <Ziegler> cd emc2.2
[02:48:02] <Ziegler> oops
[02:48:07] <Ziegler> gotta stop doing that
[02:48:09] <[Harlock]> emc2.2$
[02:48:18] <[Harlock]> heh
[02:48:55] <Ziegler> geez... now I am irc'n in the terminal
[02:50:35] <[Harlock]> OH YEAAH
[02:50:39] <[Harlock]> thanks fenn
[02:50:49] <[Harlock]> for.. ummm... your moral support
[02:51:01] <fenn> probe_parport worked?
[02:51:05] <[Harlock]> yup
[02:51:13] <[Harlock]> needs to be loaded before hal_parport
[02:51:18] <[Harlock]> at least in my case :)
[02:51:32] <[Harlock]> now I got I nice square wave on my scope
[02:51:35] <maddash> damn. My problem is that I have a three-axis system, two of which (X, Y) are controlled by nc code, while the third (Z) needs to move periodically while the other two are executing. assuming that my description is clear enough, is there any way to solve this problem without hacking into src/emc/motion/ ?
[02:52:06] <Ziegler> cant do it in g-code?
[02:52:13] <maddash> how?
[02:52:26] <Ziegler> x0y0z0
[02:52:42] <Ziegler> put all 3 movements on same line
[02:52:45] <gezr> g01 x2.0 y3.0 z-.001 more lines of x and y stuff, then another one with z-.001
[02:53:13] <Ziegler> what gezr said
[02:53:14] <cradek> you can run a motor with stepgen/freqgen/pid/etc without it hooked to emc proper
[02:53:14] <cradek> maybe that's what you're asking
[02:53:14] <cradek> just use whatever you want (not emc) for the position command
[02:53:39] <cradek> you can do things in hal that aren't hooked to emc (motion controller)
[02:53:53] <maddash> cradek: I've considered that, but barebones stepgen deprives me of the nice accel/deccel ramps
[02:54:06] <fenn> stepgen does have accel actually
[02:54:08] <cradek> there are limit hal blocks to do that
[02:54:16] <fenn> or, it did at least
[02:54:16] <cradek> or right, stepgen has them built in
[02:54:22] <Ziegler> can you give us a hint of what you are doing (I am curious )
[02:54:33] <fenn> its a silly robot
[02:55:17] <fenn> maybe i should do probe_parport on my (now busted) panel pc
[02:55:22] <maddash> Ziegler: what fenn said -- I've sort-of achieved this in pre-2.1.5, but now i've decided to take the time consider all my options before hacking into 2.2
[02:55:34] <gezr> hey fenn, I think Ill have this licked with gamma-x pretty soon, or at least he will be a lot better off
[02:55:42] <[Harlock]> fenn: maybe :)
[02:55:56] <Ziegler> sudocool maddash
[02:56:02] <Ziegler> geez doing it again
[02:56:15] <[Harlock]> fenn: what kind of I/O interface you're using?
[02:56:17] <fenn> the parport wasn't working and when i took it apart (destructively) i found that the parport connector had been re-soldered in an amateur way
[02:56:43] <fenn> sudocool, is that some sort of british/japanese game?
[02:56:43] <[Harlock]> there aren't many people soldering on mobos...
[02:57:15] <[Harlock]> ..or is it an add-on card?
[02:57:19] <Ziegler> hey... who was it here that managed to get axis running on a remote xwin server ?
[02:57:28] <fenn> it was given to me without much explanation as to why he was getting rid of it
[02:57:36] <Ziegler> lol fenn
[02:57:39] <[Harlock]> lol
[02:57:42] <maddash> cradek: there's accel ramping inside stepgen? I thought all of that was done inside tp.c
[02:57:44] <cradek> I run AXIS over ssh all the time
[02:57:46] <[Harlock]> "paperweight"
[02:57:56] <cradek> maddash: yes
[02:57:57] <Ziegler> cradek: I am trying...
[02:57:59] <maddash> Ziegler: ssh -X
[02:58:01] <fenn> well, it worked fine as a computer with touchscreen, which was neat
[02:58:07] <cradek> yeah ssh -X is the only trick
[02:58:25] <maddash> cradek: 'yes'? which one?
[02:58:28] <fenn> then i broke the touchscreen trying to get it out of the crappy plastic case
[02:58:36] <Ziegler> yes yes... I am aware of the command.. but for whatever reason... I can run the program live on the box itself... but starts having errors when I try it over ssh
[02:59:02] <[Harlock]> lol fenn
[02:59:08] <maddash> fenn: ;O
[02:59:10] <cradek> Ziegler: try saying what the error is
[02:59:13] <fenn> Ziegler: are you talking about ssh -X or using remote NML?
[02:59:41] <Ziegler> I have tried ssh, and even xdmcp remote desktop
[02:59:55] <fenn> try something simple like xclock
[03:00:07] <Ziegler> let me open up pastebin.ca
[03:00:21] <fenn> sometimes you have to enable x forwarding in ssh
[03:00:24] <maddash> Ziegler: actually, ssh -CX
[03:00:43] <[Harlock]> fenn: would you open a brand new iPhone?
[03:01:02] <[Harlock]> what is it... xhost +
[03:01:04] <[Harlock]> ?
[03:01:14] <[Harlock]> yeah
[03:01:19] <fenn> [Harlock]: hell no, i'd sell that shit on ebay
[03:01:22] <[Harlock]> that's the command: xhost +
[03:01:42] <maddash> fenn: it's not silly
[03:01:42] <cradek> xhost access is not needed when using ssh
[03:01:54] <fenn> maddash: should i cower in fear?
[03:02:42] <maddash> fenn: it'll have a laser mount soon
[03:02:59] <maddash> fenn: as soon as I get this third axis to work independent of coord mode
[03:03:28] <fenn> maybe you should run two emc's at the same time
[03:03:41] <Ziegler> http://pastebin.ca/834325
[03:04:10] <fenn> then you could mount the laser on a hexapod
[03:05:47] <Ziegler> ./configure --enable-run-in-place --enable-simulator
[03:06:07] <maddash> fenn: maybe you belong at the top of my ${HOME}/hitlist
[03:06:17] <Ziegler> thats what I used to ./configure
[03:06:39] <fenn> bring it on. my robot sentry will blow your robot to pieces
[03:07:09] <gezr> yummy
[03:07:29] <fenn> ... as soon as i find its box
[03:07:46] <gezr> I think gamma-x must have bumped his head and knocked himself out, he has disappeared from the conversation
[03:07:55] <cradek> Ziegler: _tkinter.TclError: Could not allocate font
[03:08:02] <cradek> maybe your fontpath is bad or something
[03:08:27] <[Harlock]> ok great... now I can run stepper_mm fine... but AXIS seems to have the hiccups.. it pauses for 1sec - 1.2 sec in the middle of a cut?!?!?
[03:08:46] <maddash> [Harlock]: oh noes!!!111 RT delays!! check dmesg|less
[03:08:53] <[Harlock]> lol
[03:08:56] <[Harlock]> okok
[03:09:03] <[Harlock]> I'm new to rt
[03:09:05] <[Harlock]> :P
[03:09:09] <fenn> the machine stops or the gui stops?
[03:09:34] <[Harlock]> for now the "machine" is my scope and it seems to freeze as well, but it's hard to tell
[03:10:13] <gezr> wow, not using g4 in your gcode are you?
[03:10:18] <[Harlock]> I got to admit that running latency checks for too long sometimes locks my machine
[03:10:39] <[Harlock]> it's the EMC2 AXIS letter engraving demo
[03:11:06] <gezr> how many lines is your scope watching?
[03:11:14] <[Harlock]> 4
[03:11:21] <fenn> nice scope
[03:11:35] <[Harlock]> yeah... hehe... paid few thousands for it
[03:11:36] <gezr> on seperate channels?
[03:11:41] <gezr> oh yeah then
[03:11:59] <Ziegler> cradek: yeah... it was the dang xserver on this machine
[03:12:18] <[Harlock]> I have a used Hitachi 2 channels 100Mhz that I purchased second-hand
[03:12:20] <maddash> cradek: no chance of running stepgen in sim, is there?
[03:12:28] <[Harlock]> but the flyback xfo died
[03:12:37] <[Harlock]> can't find a replacement
[03:12:46] <gezr> harlock, cpu and ram in teh machine?
[03:12:53] <[Harlock]> I was so pissed off... it was such a nice scope!
[03:13:03] <[Harlock]> but now I got a Tek TDS2024B
[03:13:11] <[Harlock]> hopefully it will last
[03:13:56] <[Harlock]> I'm all for second hand eqt for hobbying, but I got burned on this scope.
[03:16:27] <maddash> anyone? stepgen in sim?
[03:16:30] <[Harlock]> I think the digital scope lacks the feeling of the CRT
[03:17:37] <gezr> I have a faulty hp 100mhz, and an old military style tek 10mhz, they dont work very well at all, in fact they may not even work, I know one of the channels on the hp is shot
[03:18:24] <[Harlock]> I was told Tek is good for their trigger detection
[03:19:04] <[Harlock]> I was considering GW I think... lower price, more features
[03:19:22] <[Harlock]> but their must be a reason why Tek is so popular
[03:19:46] <[Harlock]> I've used Lecroy, Fluke, Yokogawa
[03:19:54] <gezr> back to your delay issue, computer has plenty of cpu and ram to go around?
[03:20:11] <fenn> maddash: sure why not
[03:20:31] <[Harlock]> gezr: shouldn't be an issue...
[03:20:55] <[Harlock]> swap isn't used when I type 'free'
[03:21:11] <[Harlock]> 65 MB left
[03:21:14] <Ziegler> lucky you
[03:21:17] <gezr> you trying to run the machine at an extremely high frequency?
[03:21:19] <[Harlock]> I mean free
[03:21:44] <gezr> I dont know what its called exactly the peroid i think
[03:21:59] <[Harlock]> it's the stepper_mm config file with hal pinout update
[03:22:14] <gezr> I dont know then
[03:22:27] <[Harlock]> I reduced the feed rate...
[03:22:34] <[Harlock]> 33%
[03:22:51] <[Harlock]> it just goes ballistic at 100%...
[03:23:18] <fenn> what's the result of your latency-test?
[03:23:22] <[Harlock]> at 100%, I get joint 0 following error
[03:23:33] <[Harlock]> uuh... lockup? ;)
[03:23:42] <fenn> huh?
[03:24:00] <[Harlock]> I had 48690 for servo thread and 21736 for base thread
[03:24:07] <[Harlock]> then it locked up
[03:24:17] <[Harlock]> I guess it's all nvidia fault
[03:24:40] <fenn> change 'nv' or 'NVidia' to 'vesa' in your xorg.conf
[03:24:57] <[Harlock]> then I don't have GL anymore, right?
[03:25:00] <fenn> right
[03:25:15] <fenn> too bad so sad
[03:25:17] <cradek> wrong, you will get software GL
[03:25:25] <cradek> perfectly fine for emc
[03:25:40] <[Harlock]> ah yes, but a pain for mplayer
[03:25:52] <cradek> doubt mplayer uses gl
[03:26:03] <[Harlock]> speed is affected for sure
[03:26:05] <cradek> ah
[03:26:15] <[Harlock]> mplayer has GL output capability
[03:26:21] <[Harlock]> two drivers at least
[03:26:24] <[Harlock]> anyhow
[03:26:35] <[Harlock]> a test is required :)
[03:26:39] <[Harlock]> bbl
[03:28:33] <maddash> brb. suckers.
[03:30:01] <[Harlock]> ok vesa up and running
[03:30:20] <[Harlock]> let's see what Axis says..
[03:30:46] <[Harlock]> doesn't workl
[03:31:31] <[Harlock]> X server has no OpenGL GLX extension
[03:31:53] <[Harlock]> ugh
[03:32:31] <cradek> you had the proprietary nvidia driver installed?
[03:32:35] <[Harlock]> Do I have to compile Mesa or something like that?
[03:32:40] <[Harlock]> yes
[03:32:45] <cradek> you have to banish it to get mesa back
[03:32:57] <cradek> I'm not sure what the trick is
[03:33:11] <cradek> but softwtare mesa does work, and it is already installed
[03:33:16] <[Harlock]> the point is "I need Mesa"
[03:33:36] <cradek> mesa is/was the default
[03:33:43] <cradek> you will get it back by getting rid of nvidia
[03:33:44] <Ziegler> is there a tool table for sim axis?
[03:33:51] <cradek> Ziegler: yes
[03:34:18] <Ziegler> in the configs/sim dir?
[03:34:23] <cradek> yes
[03:34:29] <cradek> or, File/edit tool table
[03:34:40] <Ziegler> reallllyyyy!?
[03:35:07] <Ziegler> hmm... edit is grayed out
[03:35:23] <cradek> your config does not have [DISPLAY]EDITOR defined then
[03:35:30] <cradek> just set that to your favorite editor
[03:35:34] <Ziegler> cool beans
[03:46:34] <Ziegler> does axis show tool offset ?
[03:47:22] <cradek> yes
[03:48:22] <Ziegler> t1 in my tool table is set at .5 inch
[03:48:39] <Ziegler> I issue a g42 before I begin into the movments
[03:48:45] <cradek> g43
[03:48:52] <Ziegler> but it still seems to trace the actual path
[03:49:07] <cradek> g43h1 will make the cone jump .5"
[03:49:16] <cradek> if you then g0z0, it will go back up/down to the path
[03:49:30] <Ziegler> err.. sorry meant compensation not offset
[03:50:23] <cradek> the preview shows the path of the tool
[03:50:42] <cradek> if you enable compensation, the path moves left/right
[03:51:06] <Ziegler> ok let me reload the tool table again
[03:59:49] <[Harlock]> yuk... to exorcise my xorg server, I have to dig a file out of package xserver-xorg-core_1.1.1-21etch1_i386.deg
[03:59:57] <[Harlock]> that's what I hate from Linux
[04:01:31] <cradek> what you hate from linux is that you screwed up your install wth a proprietary closed source video driver?
[04:02:24] <cradek> dangit my inside-thoughts keep leaking out
[04:02:32] <[Harlock]> lol
[04:02:53] <[Harlock]> sorry, but I would rather have everything open source
[04:03:08] <[Harlock]> but it's not the case
[04:03:27] <cradek> well don't blame linux folks for that
[04:03:31] <cradek> that's all I'm saying
[04:03:37] <cradek> it makes no sense
[04:04:03] <[Harlock]> Oh I had some fun trying to compile RTAI as well
[04:04:12] <cradek> well THAT you can complain about
[04:04:16] <[Harlock]> but it must be Nvidia fault anyway
[04:04:17] <cradek> it's a pain in the ass
[04:04:38] <[Harlock]> it's always easy to point the sealed up box
[04:04:45] <[Harlock]> "
[04:04:47] <cradek> brb
[04:06:30] <[Harlock]> I don't think the real problem is Nvidia or their closed drivers
[04:06:50] <Ziegler> alright... would this show off radius comp in axis:
[04:06:55] <Ziegler> g0 x0y0z0
[04:06:57] <Ziegler> T1
[04:06:59] <Ziegler> G42 G1x1y10 F10
[04:07:00] <Ziegler> m30
[04:07:21] <cradek> no, the tool is not loaded.
[04:07:28] <cradek> t1m6, or g42d1
[04:08:18] <Ziegler> bahhh
[04:08:20] <Ziegler> duh
[04:08:27] <Ziegler> thanks cradek
[04:08:59] <[Harlock]> bbl
[04:09:22] <cradek> beware you will have to make a proper compensation entry move
[04:09:40] <cradek> g0x0y0
[04:09:41] <cradek> g42d1
[04:09:46] <Ziegler> right I hear ya
[04:09:50] <cradek> g0x1y1
[04:09:50] <cradek> y10
[04:09:56] <cradek> ^ try those 4 lines
[04:10:15] <cradek> it will end up at x1.5 y10
[04:10:20] <cradek> x1.25 y10
[04:11:01] <cradek> and no I didn't test it :-)
[04:11:26] <Ziegler> will g42d1 do a tool change if not already using d1?
[04:11:42] <cradek> no
[04:11:49] <cradek> it only loads the radius of tool 1
[04:12:00] <cradek> to change tools you must use m6
[04:12:10] <[Harlock]> ok glxgears work
[04:12:14] <cradek> the loaded tool and radius may be from different lines in the tool table
[04:12:43] <Ziegler> oh?
[04:12:43] <[Harlock]> it's smooth for the first second then its crappy
[04:15:35] <[Harlock]> it's insane.. glxgears maxes out the cpu... with very poor refresh rate (2fps?)
[04:16:04] <cradek> yuck
[04:16:20] <cradek> sometimes glxgears lies
[04:16:22] <cradek> try emc
[04:16:36] <[Harlock]> it's loaded... i'm waiting for my scope to boot
[04:16:38] <cradek> or another gl screensaver or something
[04:17:06] <[Harlock]> is there a parameter to glxgears to display fps?
[04:17:27] <cradek> dunno
[04:17:43] <[Harlock]> "joint 0 following error"
[04:18:11] <[Harlock]> that is after letter E of outlining the demo EMC2 AXIS
[04:18:18] <cradek> what config?
[04:18:30] <[Harlock]> stepper_mm
[04:18:40] <cradek> hmm
[04:18:43] <[Harlock]> I think I have to tune something with my latency don't I?
[04:19:09] <cradek> did you edit stepper_mm at all?
[04:19:20] <[Harlock]> yeah the pinout in hal
[04:19:21] <cradek> this is a sign of misconfiguration of vel/acc constraints
[04:19:31] <[Harlock]> I didn't touch this
[04:19:41] <[Harlock]> hey my latency is waaaaay down
[04:19:47] <[Harlock]> I mean jitter
[04:20:09] <cradek> gotta run - goodnight
[04:20:13] <[Harlock]> cya
[04:20:15] <cradek> maybe someone else can help with the error
[04:20:16] <gezr> night cradek
[04:20:39] <gezr> harlock its starting to sound like you need to tweak the configs to suit your computer
[04:20:48] <gezr> I do not know how to do this though
[04:21:16] <[Harlock]> my latency jitter was way down, but it just jumped up
[04:21:33] <[Harlock]> it's at least 2 times lower on the servo thread
[04:21:49] <[Harlock]> base thread is same as with the nvidia driver
[04:22:47] <[Harlock]> I'll tune that tomorrow
[04:22:49] <[Harlock]> cya guys
[04:24:05] <Ziegler> cutter gouging with cutter radious comp
[04:25:37] <Ziegler> http://pastebin.ca/834387
[04:26:17] <gezr> what is the value of d1?
[04:26:28] <Ziegler> .5"
[04:26:45] <gezr> tool radius is greater then the comanded moves
[04:28:00] <gezr> try making d1 .075
[04:28:01] <Ziegler> I guess I dont understand cutter comp like I thought I did
[04:28:58] <gezr> I still dont fully understand it, but see what .075 does
[04:29:23] <Ziegler> I thought I could take a .2 x .2 square of gcode movment, and cut it out with a .5" cutter compensated to the left
[04:29:28] <Ziegler> trying it gezr
[04:29:55] <fenn> Ziegler: cutter comp is lame, you can't do inside corners, they have to be arcs
[04:30:29] <gezr> wait a moment Ziegler, try finishing your Z moves then call comp
[04:30:51] <gezr> and your program needs g42
[04:31:05] <gezr> your first move puts the cutter on the left of the part or right compensation
[04:31:06] <Ziegler> (tried both for the heck of it)(
[04:31:33] <Ziegler> actually I want the cutter on the left of the part
[04:32:08] <gezr> yeah your box is all backwards
[04:32:51] <gezr> you start 0, then you move right to left / then you move back / up
[04:33:02] <Ziegler> if I am going CCW around the outside of a box... I would use G41... correct?
[04:33:09] <gezr> yep
[04:33:29] <gezr> but your running it over
[04:34:24] <fenn> maybe the lead-in move should be an arc?
[04:34:36] <gezr> try this, x0 y0, y-1, x-.3, y0, x0
[04:35:07] <gezr> if you plot your points out, your tool is destroing your box in the first move
[04:36:18] <gezr> fenn : you check that link he posted/
[04:36:53] <Ziegler> http://pastebin.ca/834401
[04:37:34] <gezr> did that work?
[04:37:39] <Ziegler> nope
[04:37:41] <Ziegler> hehe
[04:37:43] <gezr> you left your d out
[04:38:04] <fenn> need a lead-in move for cutter comp, even for a stupid box
[04:38:04] <Ziegler> docs say it isnt needed... but I will add it in anyway
[04:38:22] <gezr> and use g42
[04:38:31] <gezr> your cutter is on the right hand side of the movement
[04:38:54] <gezr> cause your first move is from right to left
[04:39:02] <Ziegler> which would be clock wise
[04:39:08] <gezr> no, counter
[04:39:19] <Ziegler> 0 to -.2 is right to left
[04:39:26] <gezr> x0 -> x-.2 is a left ward move from right toleft
[04:40:22] <gezr> heh, we said it right, but yes, ccw movement
[04:40:47] <Ziegler> ok... I got ya
[04:41:03] <gezr> x0 - x.2 - y-.2 - x0 - y0 is clockwise
[04:41:38] <Ziegler> so fen... how do I do a lead in?
[04:41:50] <fenn> Ziegler: http://pastebin.ca/834406
[04:41:56] <fenn> see how the first move ramps out
[04:42:38] <fenn> well, instead you put a move some distance > r away from your path
[04:44:33] <gezr> g1x.5y.5; z0f30.;g42d1;g1x0y0;x-.2;y-.2;x0;y0;g1x.5y.5;g0 g40 z1.
[04:45:48] <Ziegler> ok... I see
[04:45:59] <gezr> did you try that and it work?
[04:46:17] <Ziegler> fenn's did... didnt try yours yet
[04:46:39] <gezr> I didnt see fenn's hmm
[04:46:41] <Ziegler> fenn you are right radius comp is lame
[04:48:46] <gezr> its just frustrating
[04:48:56] <gezr> and it wont do curves?
[04:49:41] <Ziegler> http://pastebin.ca/834406 << this is fenn's
[04:50:33] <gezr> ah
[04:51:14] <Ziegler> is this an emc thing, or a industry thing
[04:51:33] <gezr> industry, tool comp is the work of the devil
[04:51:39] <Ziegler> lol
[04:52:02] <gezr> there is a method to the madness, in 15 years I have yet to master it
[04:52:19] <gezr> sometimes I get it right, sometimes I want to gouge my eyes out
[04:53:14] <gezr> it was supposed to be a way to program your part, and the machine offsets the tool to the correct position
[04:53:16] <Ziegler> until now I have just used the cam software to generate a line that is compensated out
[04:53:38] <gezr> but it doesnt use G41 or G42 does it, it just moves the points that much out
[04:53:40] <gezr> right?
[04:53:45] <gezr> or in
[04:53:48] <Ziegler> but it is starting to become a pain every time I decided to use a dif bit
[04:54:32] <Ziegler> no it just moves the point out
[04:55:42] <gezr> http://www.timgoldstein.com/cad_cam/diacomp.htm
[04:55:55] <Ziegler> hehe... I was reading the same thing
[04:56:01] <Ziegler> no joke
[05:09:52] <Ziegler> alright.. brain done for the evening... thanks for the help tonight
[05:12:05] <Skullworks_PGAB> ug - I hate spindle center programs
[05:14:43] <Skullworks_PGAB> program for part edge geometry and put the true tool geometry in the tool table
[05:15:50] <Skullworks_PGAB> makes using regrinds or other off size tooling easy - without cutter comp errors.
[05:24:42] <gezr> i believe my file transfer is about to be complete, leaving me with a 20g hard drive to use as my new emc drive :)
[06:09:29] <Skullworks_PGAB> * Skullworks_PGAB is retiring to a warm bed with a good book.
[06:35:28] <Unit41> tig cnc ?
[06:38:04] <Unit41> http://www.cnczone.com/forums/showthread.php?t=21978
[07:08:45] <Unit41> http://www.youtube.com/watch?v=hhq1ns8Gz54
[11:04:30] <gnix_oa1> helle
[11:28:50] <anonimasu> Unit41: it's a nice project
[11:29:22] <anonimasu> but, it dosent really weld
[11:29:24] <anonimasu> http://www.cnczone.com/forums/attachment.php?attachmentid=25758&stc=1&d=1163741571
[11:29:30] <anonimasu> that's how real welds look..
[11:29:31] <anonimasu> :)
[12:41:54] <alex_joni> anonimasu: that's an ok weld :P
[13:13:13] <maddash> why is the cvs so SLLOOOOOOOOOOOOOOOOOW?
[13:13:46] <maddash> (rhetorical)
[13:25:24] <alex_joni> maddash: because you don't check out entire trees every minute or so..
[13:25:33] <alex_joni> and for regular operations it's quite fast
[13:26:10] <alex_joni> I suspect for a new checkout, downloading a tar from teh viewcvs server is faster
[13:26:18] <maddash> alex_joni: no, I was downloading emc2/ (tag=RELEASE2-2-2-BRANCH) as a tarball
[13:26:21] <alex_joni> and you can even update it lateron... (didn't try it though)
[13:26:31] <alex_joni> maddash: then it's probably the connection :)
[13:27:37] <maddash> argh
[13:28:13] <alex_joni> maddash: I think we could accept sponsorship for faster connections though
[13:29:08] <maddash> alex_joni: ;)
[13:44:30] <maddash> "EMC_TRAJ_SET_FO_ENABLE"? what if I'm at traj-set-scale(.1), and then I disabled FO?
[13:46:30] <maddash> what's with the massive disconnects?
[13:46:36] <jepler> it's only you, dude
[13:47:00] <maddash> is this place hosted on a private server?
[13:47:21] <maddash> damn, I bet it's the firewall rules
[13:47:33] <jepler> "You've reached freenode, a service of Peer-Directed Projects Center (PDPC). Freenode provides discussion facilities for the Free and Open Source Software communities, for not-for-profit organizations and for related communities and organizations. In 1998, the network had about 200 users and less than 20 channels. The network currently has nearly 40,000 users" http://freenode.net/
[13:48:06] <maddash> that's odd, because I'm on #debian also, and I haven't dropped at all
[13:48:26] <maddash> no, wait, I was wrong.
[13:48:40] <maddash> I'm getting shafted by the network admin
[13:49:20] <jepler> maddash: anything which changes the net feed scale causes a ramp up or down to the new net feed rate while staying within the acceleration constraints of the machine.
[13:50:05] <maddash> jepler: actually, I was asking about what 'set-fo-enable/disable' does
[13:50:26] <maddash> jepler: does 'set-fo-disable' instantly drop me to feedscale=0?
[13:50:54] <maddash> also, is the gigantic list inside emcFormat (emc.cc) organized alphabetically?
[13:51:27] <jepler> maddash: feed override enable changes whether the feed override is applied to calculate the net feed rate or not
[13:51:31] <jepler> it doesn't change the feed override value
[13:51:58] <jepler> if it did, though, it would be more like setting the feed override to 1.0 or 100%, not to 0.
[13:52:55] <maddash> jepler: so FR changes would be apparent if 'set-fo-disable' is issued while FO!=1.0?
[13:53:30] <jepler> yes that's my interpretation of tfs
[13:58:14] <maddash> 'tfs'? traj-feed-scale?
[14:00:57] <jepler> the f*&&&&ing source
[14:02:25] <maddash> :)
[14:02:33] <maddash> * maddash scurries off to read tfs
[14:15:58] <maddash> coool, i've stopped disconnecting
[14:28:59] <xemet> hello
[14:31:18] <xemet> jepler are you there?
[14:36:17] <alex_joni> hi xemet
[14:36:24] <xemet> hi
[14:36:54] <xemet> I've built a pwm amplifer based on L298 like the one shown on jepler page
[14:37:30] <xemet> http://emergent.unpythonic.net/projects/01142347802
[14:37:48] <xemet> I've a small servo exactly the same used by jepler
[14:37:59] <xemet> and I would like to tune the PID
[14:38:08] <xemet> suggestions?
[15:13:34] <maddash> who's the funny guy named 'KL' inside rs274_pre.cc?
[15:13:43] <maddash> er, rs274ngc_pre.cc*
[15:14:11] <alex_joni> maddash: my bet it's ken lerman
[15:14:22] <maddash> oh. haha. "!!!KL"
[15:14:49] <alex_joni> * alex_joni doesn't see anything funny about that
[15:14:54] <maddash> "!!!"
[15:14:58] <lerman> That would be me...
[15:15:02] <alex_joni> hi ken :)
[15:15:14] <alex_joni> * alex_joni still doesn't see anything funny about that
[15:15:26] <lerman> I have this practice of noting places that might need work by commenting !!!KL...
[15:15:37] <lerman> That's easy to search for.
[15:15:44] <lerman> Hi alex_joni
[15:15:48] <alex_joni> yup, I do that too sometimes
[15:15:56] <alex_joni> lerman: how's christmas over there?
[15:16:05] <maddash> lerman: unless you're writing a complex piece of irc/spam script
[15:16:16] <maddash> lerman: if the code doesn't follow protocol correctly, then why not annihilate it altogether?
[15:16:20] <lerman> Hanukah is over.... How was your holiday.
[15:16:34] <maddash> haha...owned.
[15:16:48] <alex_joni> lerman: quite good
[15:16:56] <lerman> There are things worth doing that aren't worth doing right. At least not right now.
[15:17:18] <maddash> also, I object to, "int Interp::read(const char *command) //!< may be NULL or a string to read"
[15:17:53] <lerman> My end of year has been an SOB. Had my prostate removed on Dec 5. But they got all the cancer, so that's a good point.
[15:18:29] <maddash> no, wait, I don't. "read(0)" from inside ::execute()
[15:18:59] <lerman> maddash: My general approach in modifying someone else's code is to make as few changes as possible. If you break it, you own it.
[15:20:11] <lerman> I don't generally care for all of the C++ class crapola, but I'm not about to change all the code to my style. Particluarly since the next person to touch it will change it to another style.
[15:20:36] <maddash> what 'c++ class crapola'?
[15:20:47] <alex_joni> lerman: sorry to hear that, but glad you're feeling better
[15:21:27] <lerman> It doesn't make sense to me that Interp is a class. -- after all, there will never be more than one instance of it.
[15:21:53] <maddash> word.
[15:22:05] <alex_joni> lerman: the only advantage I see, if someone decides to implement a variant, and use this as a template
[15:22:14] <alex_joni> or rather inherit from it..
[15:22:47] <cradek> it does seem like you could derive a modified interp from the base interp
[15:23:04] <maddash> **seem**
[15:23:10] <cradek> we haven't done that yet, but we will eventually need various interps
[15:23:18] <maddash> why?
[15:23:19] <lerman> Easier said than done, I suspect.
[15:23:27] <alex_joni> maddash: until proven contrary.. it should work :)
[15:23:47] <lerman> Until demonstrated it can't work. :-)
[15:23:48] <alex_joni> maddash: probably because no-one cared enough to implement a different interp
[15:23:49] <cradek> that's the attitude
[15:23:51] <maddash> alex_joni: right, I the flying spaghetti monster sent me coal for christmas
[15:24:16] <alex_joni> he did? I thought he only sent some to me
[15:24:17] <alex_joni> fscker
[15:24:57] <alex_joni> * alex_joni strikes the spaghetti monster from the easter gift list
[15:33:45] <alex_joni> "Frankly I'm suspicious about anyone who has a strong oppinion on a complicated issue"
[15:35:15] <maddash> and he has returned...
[15:38:55] <[Harlock]> does it make sense to have a BASE_PERIOD=100000 ?
[15:39:49] <[Harlock]> my servo thread jitter used to be around 48690 ns, but now it's 24283 after removing nvidia
[15:40:26] <alex_joni> that's quite large for a BASE_THREAD
[15:40:39] <alex_joni> 100usec means you only get 10k steps max
[15:40:54] <alex_joni> (using doublestep), without you get max 5k steps / second
[15:41:59] <[Harlock]> RTAI was complaining when I had 50000 with jitter of 48690
[15:42:35] <[Harlock]> but now that I've thrown out Nvidia, I get half the jitter I had before
[15:43:11] <alex_joni> [Harlock]: run a latency test, and tell us the max jitter
[15:43:25] <[Harlock]> it'r running right now
[15:43:40] <[Harlock]> servo thread (1.0ms) is at 24283 ns
[15:43:48] <alex_joni> try to abuse the system while it's running
[15:43:52] <alex_joni> glxgears
[15:43:56] <alex_joni> switching windows
[15:44:01] <[Harlock]> yeah glxgears maxes out the cpu
[15:44:02] <alex_joni> etc
[15:44:36] <[Harlock]> funny thing with glxgears is it runs smooth for 1 second then it's about 1fps
[15:45:04] <[Harlock]> the gears are jumping
[15:45:12] <[Harlock]> probably less than 0.5fps
[15:45:25] <alex_joni> [Harlock]: what cpu?
[15:45:38] <jepler> if you're using stepconf, BASE_PERIOD = 100000 is what it writes out when your max step rate is 10kHz or below.
[15:46:06] <alex_joni> depends on the machine, but 10kHz seems awfully slow
[15:46:15] <[Harlock]> can't use stepconf, it complains about missing glide or glade.
[15:46:27] <[Harlock]> (another package to dig..)
[15:46:29] <jepler> really? did you install from the package, or in some other way?
[15:46:41] <[Harlock]> compiled as per Debian Etch instrucitnos
[15:46:45] <jepler> ah
[15:46:49] <[Harlock]> RTAI, and so on
[15:47:19] <alex_joni> [Harlock]: you should be able to add the repo to apt, and apt-get build-dep emc2
[15:47:26] <alex_joni> that should install all deps you need
[15:47:31] <[Harlock]> but with kernel 2.6.20.2
[15:47:46] <[Harlock]> including kernel/
[15:47:48] <cradek> that may or may not work right. I bet many of the packages are differently named
[15:47:50] <alex_joni> (even if you don't install the packages from our repo)
[15:48:02] <jepler> the package you need is probably python-glade2 and probably python-gnome2
[15:48:17] <alex_joni> cradek: I'd use it as a starting point..
[15:48:17] <jepler> but those are ubuntu package names and as cradek says sometimes they are named differently
[15:48:38] <[Harlock]> thx jepler, i'll be installing this while the latency test is running
[15:48:59] <jepler> [Harlock]: if you've already got an (almost) working configuration, it may make you use more time than you save
[15:49:23] <[Harlock]> python-glade2 is installing
[15:50:01] <[Harlock]> jitter is at 23141 for 1.0ms thread, and 38891 for the 25.0 us thread
[15:50:30] <[Harlock]> glxgears are running, i'm doing rgrep a * from root, and installing packages
[15:51:56] <[Harlock]> the latency test ran for 19minutes, that should be good
[15:54:41] <[Harlock]> why is stepconf looking for /etc/emc2/emc2-wizard.gif ???
[15:55:58] <alex_joni> [Harlock]: did you configure for run-in-place or for installed?
[15:58:06] <[Harlock]> run-in-place
[15:58:21] <[Harlock]> as per Debian Etch install instruction
[15:58:33] <alex_joni> did you . scripts/emc-environment ?
[15:58:52] <[Harlock]> hmmm maybe not
[15:58:54] <[Harlock]> hold on
[16:00:48] <[Harlock]> do I have to add this to bash rc ?
[16:01:35] <[Harlock]> wait.. what is the leading dot for?
[16:02:01] <jepler> you have to do it before running any emc programs from an RIP install. ". filename" is the same as "source filename"; it reads it into the running shell.
[16:02:23] <[Harlock]> never saw that before
[16:02:44] <anonimasu> alex_joni: yeah but not the one in the forum :p
[16:02:54] <jepler> but I'm not sure that'll solve this problem of emc2-wizard.gif
[16:03:00] <[Harlock]> I saw export variable but not . file
[16:03:07] <[Harlock]> anyhow
[16:03:49] <[Harlock]> gobject.GError: Failed to open file '/etc/emc2/emc2-wizard.gif': No such file or directory
[16:05:54] <cradek> % . scripts/emc-environment
[16:05:58] <cradek> % strace -eopen stepconf 2>&1|grep emc2-wizard
[16:05:58] <cradek> open("/etc/emc2/emc2-wizard.gif", O_RDONLY|O_LARGEFILE) = 15
[16:06:21] <cradek> confirmed as a bug here
[16:06:27] <jepler> stepconf actually searches a sequence of directories to find emc2-wizard; it takes the first one it finds
[16:06:31] <jepler> wizard = os.path.join(datadir, "emc2-wizard.gif")
[16:06:33] <jepler> if not os.path.isfile(wizard):
[16:06:35] <jepler> wizard = os.path.join("/etc/emc2/emc2-wizard.gif")
[16:06:38] <jepler> if not os.path.isfile(wizard):
[16:06:40] <jepler> wizdir = os.path.join(os.path.abspath(os.path.dirname(__file__)), "..")
[16:06:43] <jepler> wizard = os.path.join(wizdir, "emc2-wizard.gif")
[16:07:41] <acemi> [Harlock]: there are some missing packages in the dependencies list which is in Debian Lenny instructions
[16:07:51] <[Harlock]> O_o
[16:08:27] <jepler> argh bbl
[16:08:51] <[Harlock]> this EMC is from CVS pre-2.2
[16:09:13] <[Harlock]> I've redownloaded the latest CVS, but now some RTAI tests fails
[16:09:17] <acemi> the packages to compile documentations and python-glade is missing
[16:13:09] <[Harlock]> bbl
[16:17:07] <alex_joni> RTAI tests fail?
[16:17:43] <acemi> [Harlock]: sorry, I think that you use Lenny. Etch instructions are OK
[16:19:44] <acemi> but if there is new dependencies after EMC2 2.1.7, theese may be missing
[16:20:48] <eric_U> cable guys have been at the end of the street for an hour, now I'm paranoid
[16:25:21] <maddash> wut iz dis 'adaptive feed' shizzle?
[16:27:28] <eric_U> so now that I've upgraded to 2.2, the package manager no longer tells me that emc updates are available
[16:30:57] <eric_U> anyone know the command to restore sane properties to a terminal?
[16:31:09] <alex_joni> maddash: what does it sound like?
[16:32:52] <maddash> alex_joni: it sounds like, "AH-dap-tif FEE-d'
[16:33:05] <maddash> wtf is it adapting to?
[16:33:37] <eric_U> ok, so it's "stty sane"
[16:35:41] <jepler> maddash: the value on the HAL pin motion.adaptive-feed. This value is another value that can contribute to the net feed, separate from feed override.
[16:35:58] <jepler> possible uses for adaptive feed are EDM gap voltage sensing or spindle load sensing
[16:36:14] <jepler> http://linuxcnc.org/docs/2.2/html/gcode_main.html#sec:M52:-Adaptive-Feed-Control
[16:41:07] <Ziegler> anyone got the patience to help me out with some cutter comp g-code ?
[16:42:03] <maddash> so if I hook up a motion sensor to my bot, then adaptive feed would provide a flight-or-fight response. cool.
[16:42:05] <acemi> beles set bulmusken liseyi yeniden mi okusak acaba
[16:42:14] <acemi> beles set bulmusken liseyi yeniden mi okusak acababeles set bulmusken liseyi yeniden mi okusak acaba
[16:42:15] <acemi> beles set bulmusken liseyi yeniden mi okusak acaba
[16:42:22] <acemi> sorry
[16:42:25] <Ziegler> http://pastebin.ca/834968
[16:42:54] <acemi> these are for another channel
[16:43:32] <Ziegler> line 721 starts the beginning
[16:44:25] <alex_joni> acemi: what was that?
[16:44:32] <alex_joni> sounds like turkish to me :)
[16:45:29] <acemi> yes turkish
[16:45:38] <alex_joni> you speak it?
[16:45:50] <acemi> yes, ı!m living in istanbul
[16:48:02] <alex_joni> acemi: really? wow.. I always thought you're in the states
[16:48:19] <acemi> :)
[16:48:26] <alex_joni> * alex_joni wonders why...
[16:49:00] <alex_joni> hmm.. I always wanted to visit istanbul, but never got around to do it :)
[16:49:00] <acemi> my english is very bad, why do you think like this
[16:49:37] <maddash> WHO JUST CTCP'ed me?
[16:49:37] <alex_joni> acemi: I think it's fine.. and the work on the wiki proves it :P
[16:49:59] <alex_joni> maddash: not sure about you, but for me it was freenode
[16:50:09] <jmkasunich> acemi: many people in the states who speak english as their only language still "sound" bad when they type in IRC
[16:50:16] <alex_joni> 18:29 freenode-connect [freenode@freenode/bot/connect] requested CTCP VERSION
[16:50:17] <jmkasunich> because they can't spell
[16:50:29] <maddash> alex_joni: oh. I thought "freenode-connect" was just a blanket masquerade for some mischief-maker.
[16:52:16] <alex_joni> hi jmkasunich
[16:52:42] <jmkasunich> hi alex
[16:52:59] <maddash> lo jmkasunich
[16:53:12] <maddash> * maddash <--- mischief-maker
[16:54:03] <eric_U> do via chipsets have latency probs?
[16:54:26] <maddash> sweeeeeeeeet. my robot can now turn its head and walk at the same time.
[16:55:16] <maddash> eric_U: http://issaris.org/rtai/list.php?offset=328&runtime=&sort_order=rtai_up_latency_max
[16:55:24] <maddash> eric_U: from http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Hardware_Requirements
[16:58:53] <cradek> Ziegler: do you have a specific question about cutter compensation?
[17:03:03] <cradek> hmm my connection is terrible today.
[17:04:08] <eric_U> what's the emc latency test called?
[17:04:31] <cradek> latency-test
[17:04:31] <eric_U> rtai latency test gives me about 20 microsecond latency
[17:21:15] <BigJohnT> alex_joni: yesterda you said that EMC moves the Z axis when commaned by the external THC. How is the up/down linked to the Z axis?
[17:23:52] <alex_joni> BigJohnT: basicly it depends how you set it up, but the most common way of doing it is using external THC and up/wodn signals to emc2
[17:25:47] <alex_joni> there is a config included with emc2
[17:26:16] <BigJohnT> what do you connect the signals to in EMC?
[17:26:32] <BigJohnT> I'm at my machine shop so I can't look at EMC
[17:26:58] <jmkasunich> lerman: O-words are cool!
[17:27:18] <alex_joni> it has some documentation in openoffice (some big diagram of the HAL components Dallur used to accomplish this)
[17:27:21] <jmkasunich> my lathe is cutting a keyway as I type
[17:27:30] <BigJohnT> do you send pulse and direction to EMC or just a up/down
[17:27:45] <BigJohnT> I have that at my fab shop I'll look again at it
[17:27:57] <BigJohnT> it's a real big diagram!
[17:29:29] <alex_joni> BigJohnT: it's not a trivial setup :)
[17:29:52] <BigJohnT> makes it more challenging :)
[17:32:53] <jepler> stepgen in "v" mode (or, in older systems, freqgen) can generate steps in a direction and speed specifed by the input velocity. So if you have two signals coming in to emc: "move torch up", "move torch down", you use another component such as mux4 to feed a Z velocity of 0.0 into the stepgen if both are FALSE; +1.0 if "up" is TRUE; -1.0 if "down" is TRUE. stepgen will impose an acceleration ramp (up and down). You'll have to fine-tune the acceleration
[17:35:55] <alex_joni> jepler: but then you lose using the Z axis as a positioning axis
[17:36:02] <jepler> yeah true
[17:37:27] <alex_joni> Dallur did what you said, but using an adder, so there is a Z-pos, fb in addition to the THC stuff
[17:46:31] <Ziegler> cradek: yes... what causes the error: Cutter gouging with cutter radius comp
[17:47:03] <cradek> did you read the radius comp section of the users guide?
[17:47:10] <jepler> http://linuxcnc.org/docs/2.2/html/gcode_main.html#r9_4_1_7
[17:47:30] <cradek> it means the cutter has already cut into the edge you wanted to be next to
[17:48:28] <Ziegler> oh.. .I havn't come across that link yet
[17:48:32] <cradek> you need to program a correct entry move. you can't just turn comp on willy-nilly
[17:48:32] <cradek> thanks jepler
[17:50:11] <Ziegler> yes fenn mentioned that last night. I am just having a bit of trouble getting the hang of it yet
[17:50:26] <Ziegler> this doc helps quite a bit too
[17:50:26] <cradek> If the tool is already covering up the next XY destination point when cutter radius compensation is turned on, the gouging message is given when the line of NC code which gives the point is reached. In this situation, the tool is already cutting into material it should not cut.
[17:50:26] <cradek> ^ from jepler's url
[17:50:27] <cradek> there are a couple example files in your distribution
[17:50:31] <cradek> I think they are named comp-something
[17:50:47] <Ziegler> yes, found those too
[17:50:52] <cradek> start with a simple program - like cutting around the outside of a square
[17:51:27] <Ziegler> that is a good idea
[17:51:29] <cradek> I can help with that :-)
[17:51:36] <cradek> but your other program is too long to understand
[17:53:50] <Ziegler> seems I have a corner that will not fit also
[17:54:04] <Ziegler> interpreter will not ignore that will it
[17:54:29] <cradek> nope the tool has to fit everywhere along the path when you use cutter comp
[17:55:40] <cradek> there is no provision for it coming off the path where it can't touch it
[17:58:29] <cradek> in practical terms that only means that inside corners must be arcs with radius greater than the tool radius
[17:58:30] <cradek> you don't have to be careful about outside corners
[18:02:09] <Ziegler> that makes sense
[18:03:01] <cradek> that has the dubious benefit of the programmed path actually matching the final part
[18:03:01] <cradek> but, it's generally just a pain
[18:07:00] <Ziegler> once I get the hang of it, it going to save me a bit of time
[18:07:10] <Ziegler> but its going to take me a bit of time to get used to it
[18:10:59] <Ziegler> bbl
[18:12:41] <BigJohnT> jepler: thanks for the info on THC, just got bact to the shop
[18:15:13] <eric_U> interesting, I've been running latency tests and glxgears for a while over ssh, and the max jitter turns out to be the same as if I run it locally
[18:16:20] <lerneaen_hydra> oh, random thing I noticed with 2.2.x, if I set the base period to something too large for the max stepping speed to be acheivable then EMC won't warn me like it did in 2.1.x
[18:16:35] <lerneaen_hydra> leading to what could be something that's hard to debug
[18:17:51] <eric_U> that could be bad
[18:31:05] <BigJohnT> Dallur: on your torch height control did you use just an up and down bit from the THC box to raise and lower the Z?
[18:37:55] <xemet> hi
[18:38:34] <xemet> is there a conventional method to tune PID?
[18:39:36] <anonimasu> xemet: conventional?
[18:39:40] <anonimasu> usually you use a scope
[18:39:47] <xemet> ok
[18:39:53] <xemet> can you indicate how?
[18:40:15] <anonimasu> xemet: there's a description on how you do it in the wiki I think
[18:40:16] <xemet> http://emergent.unpythonic.net/projects/01142347802
[18:40:33] <anonimasu> it's the same thing when using halscope
[18:40:36] <xemet> I've built a PWM amplifer based on L298, similar to that made by jepler
[18:41:01] <xemet> I've a tiny motor, the same used by jepler
[18:41:18] <xemet> now I would like to get it work
[18:41:29] <anonimasu> Well, read the wiki on how to tune a pid loop :)
[18:41:36] <anonimasu> well your pid loop..
[18:41:56] <anonimasu> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Tuning_EMC2/HAL_PID_Loops
[18:43:36] <xemet> well, the first 4 steps are ok
[18:43:50] <anonimasu> http://www.expertune.com/tutor.html
[18:46:07] <Dallur> BigJohnT: Yes
[18:46:56] <Dallur> BigJohnT: Sorry if I'm a bit slow to respond, been on the road for the last couple of days
[18:47:20] <BigJohnT> Dallur: I'm d/l openoffice for windows so I can print out you chart in D size on my plotter
[18:47:38] <BigJohnT> I'm back and forth running a machine
[18:48:00] <Dallur> BigJohnT: great, is it readable at that size ? I never viewed it on anything other than a monitor
[18:48:21] <BigJohnT> Dallur: yes it will be 24" x 36"
[18:48:29] <BigJohnT> even my eyes can see that
[18:48:55] <BigJohnT> Dallur: did you loose control from EMC for the Z axis?
[18:50:10] <Dallur> BigJohnT: basically the Z axis can be controlled from EMC but when the torch is turned on the Z axis is disconnected from the motion controller and conected to a realtime loop which looks at the up/down bits
[18:50:35] <BigJohnT> Dallur: ok that makes sense
[18:50:52] <Dallur> BigJohnT: This is then integrated with a corner height lock
[18:51:11] <Dallur> BigJohnT: to prevent the torch from dipping down when slowing down at corners
[18:51:32] <BigJohnT> Dallur: corner height lock is in the classicladder or the hal?
[18:51:40] <alex_joni> whooo.. this sooo awesome :)
[18:51:46] <alex_joni> * alex_joni is uploading a movie now
[18:52:15] <BigJohnT> alex_joni: what movie is that?
[18:52:19] <Dallur> BigJohnT: Depends on which config you are using, my first plasma table config called dallur-thc uses classicladder and does not have CHL
[18:52:47] <alex_joni> BigJohnT: you'll see soon :)
[18:52:53] <Dallur> BigJohnT: the second config called plasma-thc is HAL only and has a corner height lock routine
[18:52:55] <alex_joni> something very nice though
[18:53:23] <BigJohnT> Dallur: ok I'll look for both and study them thanks
[18:53:25] <alex_joni> it'll take quite some while.. (my connection is not the best atm)
[18:57:04] <alex_joni> 5%
[18:57:16] <Dallur> BigJohnT: are you using a tip voltage measurement to generate up/down signals ?
[18:59:05] <xemet> What's FF1?
[18:59:14] <jmkasunich> feedforward
[18:59:16] <alex_joni> feed-forward 1
[18:59:24] <xemet> how I use it?
[18:59:26] <alex_joni> the derivative of FF0
[18:59:34] <xemet> when should I use it?
[18:59:39] <jmkasunich> FF0 produces an output proportional to the command position, FF1 pro
[18:59:41] <jmkasunich> oops
[19:00:06] <alex_joni> xemet: you can tune the motor using only PID, without touching FFx.. some people say that FF0=1 should be a good value to use (I think)
[19:00:09] <jmkasunich> FF0 produces an output proportional to the command position, FF1 produces an output proportional to command velocity, and FF2 produces an output proportional to command accel
[19:00:46] <xemet> an output?
[19:00:49] <alex_joni> FFx are terms whcih work with the change of command, not with the accumulated error like PID
[19:00:54] <lerneaen_hydra> is there an FF3 or higher?
[19:01:07] <lerneaen_hydra> if you want to get jerk or it's derivatives
[19:01:09] <jmkasunich> if everything is scaled right, and PID command is position, and PID output is velocity, in the same units, then FF1 = 1 will improve PID tuning
[19:01:20] <jmkasunich> lerneaen_hydra: no, although they could be added
[19:01:33] <alex_joni> xemet: the PID controller takes 2 inputs (commanded position, and feedback), and computes the next output (command that goes to the motor)
[19:01:34] <jmkasunich> problem is that the more you differentiate, the more you amplify quantization noise
[19:01:41] <lerneaen_hydra> oh, ok. How long is the sample base?
[19:01:45] <lerneaen_hydra> the dt
[19:02:01] <lerneaen_hydra> yeah exactly
[19:02:02] <alex_joni> depends on the thread you run it in..
[19:02:07] <jmkasunich> whatever the thread period is - most people are doing PID in the servo thread, and most people are running 1mS servo threads
[19:02:17] <lerneaen_hydra> oh so the servo thread
[19:02:24] <xemet> I've 15000 and 500000
[19:02:28] <xemet> base and servo
[19:02:41] <alex_joni> you could run it faster, but it makes senseonly if you can get fb faster
[19:02:49] <lerneaen_hydra> so that would be in the range of 1000-10000 counts/period?
[19:03:02] <alex_joni> xemet: why did you change servo thread?
[19:03:11] <lerneaen_hydra> (counts from the encoder)
[19:03:23] <alex_joni> lerneaen_hydra: that depends on the machine..
[19:03:30] <xemet> well. don't remember
[19:03:57] <lerneaen_hydra> yeah of course, in general though (assuming there is a "in general")
[19:04:25] <xemet> ok, I will try to understand
[19:05:02] <alex_joni> xemet: did you understand what the PID terms mean?
[19:05:15] <xemet> my encoder has 100 counts/rev, there is a 1:6 gearbox on the shaft, so I've 600 counts for one rev
[19:05:19] <xemet> alex, yes
[19:05:29] <xemet> Proportional Integrative, derivative
[19:05:41] <xemet> the problem is, where do I start?
[19:05:44] <xemet> I set P?
[19:05:51] <xemet> 1, 10, 100, 1000?
[19:06:00] <xemet> how can I choose a value
[19:06:08] <jmkasunich> sigh
[19:06:22] <xemet> ok
[19:06:25] <jmkasunich> what inits are you using, inches or mm?
[19:06:38] <xemet> I'm not able to do it that's ovious
[19:06:41] <jmkasunich> (machine units)
[19:06:43] <xemet> but I will learn
[19:06:46] <xemet> mm
[19:06:48] <jmkasunich> what inits are you using, inches or mm?
[19:06:55] <jmkasunich> ok
[19:07:05] <lerneaen_hydra> looks like nobody saw what I said before; "random thing I noticed with 2.2.x, if I set the base period to something too large for the max stepping speed to be acheivable then EMC won't warn me like it did in 2.1.x"
[19:07:13] <jmkasunich> what is your feedback scale - how many counts per mm?
[19:07:21] <xemet> 600
[19:07:33] <xemet> but, if I've the encoder in x4 mode?
[19:07:47] <xemet> are they 2400? Or not?
[19:07:54] <jmkasunich> should be
[19:07:57] <xemet> ok
[19:07:57] <BigJohnT> Dallur: yes I plan on using the arc voltage to control up/down if possible
[19:07:59] <jmkasunich> you need to figure that out
[19:08:01] <xemet> so 2400
[19:08:01] <alex_joni> lerneaen_hydra: I noticed..
[19:08:04] <xemet> for a mm
[19:08:11] <xemet> that's ok
[19:08:13] <xemet> I tried
[19:08:32] <jmkasunich> xemet: can you move your machine manually (is it a small machine with cranks or a big one?)
[19:08:33] <alex_joni> xemet: first make sure it all is defined as it should be
[19:08:36] <alex_joni> before you start tuning
[19:08:57] <alex_joni> set the proper scale, then move the machine by hand 10mm
[19:09:03] <Dallur> BigJohnT: buying a ready made one or making one ?
[19:09:05] <xemet> jm: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Tuning_EMC2/HAL_PID_Loops
[19:09:08] <alex_joni> and see if the display moved 10mm
[19:09:14] <xemet> the first four steps are ok
[19:09:23] <lerneaen_hydra> alex_joni, hmm, ok, do you know anything about it?
[19:09:34] <jmkasunich> xemet: then why are you asking about the x4?
[19:09:37] <alex_joni> lerneaen_hydra: nope.. try running from a terminal
[19:09:51] <xemet> to be sure
[19:09:55] <alex_joni> or look in dmesg to see if there's anything there
[19:09:59] <jmkasunich> if the scale is right (tested in step 4) then there is no question
[19:10:19] <xemet> ok, stupid question.
[19:10:22] <jmkasunich> there is nothing I can tell you from 1000 miles away that is more "sure" than what you can see with your own eyes on your own machine
[19:10:26] <lerneaen_hydra> alex_joni, alright, I'll test that next time I run the mill
[19:11:26] <alex_joni> * alex_joni thinks jmkasunich is on an island in the pacific
[19:11:30] <jmkasunich> xemet - so you are at step 5? velocity or torque mode?
[19:11:42] <jepler> lerneaen_hydra: for some reason jmkasunich was unhappy with the presence of the prompt, so at the time the compromise was that it would stay in 2.1 but we'd have to come up with a "better" solution for 2.2. Well, 2.2 arrived and nobody figured out anything smarter to do, so it ended up being shipped without the warning either.
[19:11:43] <jmkasunich> alex_joni: >1000, ok?
[19:11:47] <xemet> I think velocity
[19:11:50] <jepler> lerneaen_hydra: (about the step rate)
[19:12:13] <lerneaen_hydra> jepler, hmm, I see. what was wrong with a prompt?
[19:12:35] <jmkasunich> my issue was that stepgen is a generic component and shouldn't be sending messages thru an EMC specific channel
[19:12:44] <jmkasunich> I want to be able to use stepgen for other things
[19:12:51] <alex_joni> jmkasunich: oh, too bad.. an island would have been nicer
[19:12:53] <cradek_> huh, I thought I got that warning very recently on trunk
[19:13:07] <lerneaen_hydra> oh internal matters more than the concept of an error window displaying that the requested step rate it too high?
[19:13:27] <jmkasunich> mostly
[19:14:13] <BigJohnT> Dallur: I have not come to a conclusion yet on the THC
[19:14:19] <jmkasunich> stepgen already refuses (refused?) to accept a setting higher than it could generate - if you tried to set max_vel too high, it would set it at whatever the max actually is
[19:14:25] <cradek_> I said back then that it's important to have that error in the emc2 releases. My opinion has not changed.
[19:14:28] <jepler> http://cvs.linuxcnc.org/cvs/emc2/src/hal/components/stepgen.c.diff?r1=1.45;r2=1.46 -- I'm pretty sure it was removed on TRUNK in jan 2007 and never put back
[19:14:31] <cradek_> cradek_ is now known as cradek
[19:14:35] <jmkasunich> but it didn't wave a big flag in your face to announce that
[19:15:55] <lerneaen_hydra> jmkasunich, hmm, that's not the effect I noticed
[19:16:03] <jmkasunich> what did you notice?
[19:16:48] <alex_joni> lerneaen_hydra: the effect will be that stepgen will be limited, and emc2 will command faster moves, and you'll end up with following errors
[19:17:19] <lerneaen_hydra> if I set the base period to 10x my "safe" speed I noticed that when jogging the motors would sound very strange, similar to what it sounds like when a hardware travel limit stops the table, a sort of growling. the table moved, though very slowly
[19:17:22] <jmkasunich> xemet: sorry, too many conversations at once
[19:17:30] <lerneaen_hydra> unfortunately I didn't attach a scope
[19:17:46] <lerneaen_hydra> it didn't sound at all like a graceful reduction in speed
[19:17:46] <jmkasunich> lerneaen_hydra: I'm not understanding what you are saying
[19:17:56] <jmkasunich> base period 10x safe speed <--- not parsing
[19:17:56] <xemet> don't worry
[19:18:07] <jmkasunich> 10 times the period? or 10 times the frequency
[19:18:12] <jmkasunich> and what is safe speed?
[19:18:14] <xemet> I will be back later
[19:18:31] <jmkasunich> ok
[19:18:47] <lerneaen_hydra> uh, I increased my base period to 10 times my previous, known safe period (stepgen was cabable of generating steps up to maxvel)
[19:19:09] <jmkasunich> making base period 10 times longer is never a good thing IMO
[19:19:22] <lerneaen_hydra> ie I went from 40µs to 400µs
[19:19:26] <jmkasunich> you should run the base period as fast as your system will let you, to minimize step jitter
[19:19:32] <maddash> the call structures of emctaskissuecommand are quite inefficient
[19:19:46] <jmkasunich> 400uS is horrible jitter
[19:19:54] <jmkasunich> no wonder your motors sound unhappy
[19:20:05] <alex_joni> maddash: last time you started to get quite efficient with diff & pastebin
[19:20:06] <lerneaen_hydra> so the jitter caused of the noise?
[19:20:11] <alex_joni> maddash: last time you started to get quite efficient with diff & pastebin
[19:20:19] <jmkasunich> lerneaen_hydra: probably
[19:20:28] <jmkasunich> why did you make base period longer?
[19:20:32] <lerneaen_hydra> jmkasunich, oh ok, good
[19:20:38] <jmkasunich> (unless you have a very good reason, don't do that)
[19:20:53] <alex_joni> and even if you do.. don't do that :P
[19:21:11] <maddash> alex_joni: is that code for a patch request?
[19:21:30] <jmkasunich> well, if the reason is "I'm running a P100, and I only need 100 steps per second", then 400uS is probably OK
[19:21:50] <alex_joni> maddash: and you're picking up subtle hints more ferquently :D
[19:22:08] <lerneaen_hydra> I kept increasing it because I got strange sporadic unexpected RT errors (latency test showed a max ovl of 30µs, my set base period was 40µs. as I increased from 40µs which gave error messages after a minute or so the messages dissapeared when base period was set to around 50-60µs
[19:22:24] <alex_joni> http://www.youtube.com/watch?v=TLxwAX8G3oI - yay emc2
[19:22:35] <maddash> alex_joni: I'd be happy to submit a patch, but I want to make sure that we all agree first
[19:22:40] <jmkasunich> lerneaen_hydra: then 60uS is the number to use, not 400uS ;-)
[19:23:00] <alex_joni> maddash: I don't think we'll agree
[19:23:17] <alex_joni> developers around emc2 each have their area of interest
[19:23:22] <lerneaen_hydra> yeah, I was increasing it a lot more becuase I was expecting a prompt with an "unacheivable step rate" window, and never got one
[19:23:30] <alex_joni> in that area when they see something not good enough, they tend to improve it
[19:23:39] <maddash> brb, I'm gonna shoot my net admin in the foot
[19:23:52] <maddash> then, I'll talk about taskissuecommand
[19:23:53] <jmkasunich> alex_joni: who built the robot? you?
[19:23:54] <lerneaen_hydra> alex_joni, cool :)
[19:24:11] <lerneaen_hydra> alex_joni, so who did you do favors for to get the robot?
[19:24:13] <alex_joni> jmkasunich: nope, a guy from germany I've helped through a forum
[19:24:27] <alex_joni> and a "couple" of emails
[19:25:17] <alex_joni> I just got another movie from him via email, but it takes close to an hour for me to upload..
[19:29:47] <maddash> 14:22:33 <maddash> alex_joni: I'd be happy to submit a patch, but I want to make sure that we all agree first SAS BY (1)
[19:29:59] <maddash> err, whoops
[19:30:08] <alex_joni> sas by (1)?
[19:30:48] <maddash> alex_joni: a fragment of a seperate message combined with the effects of the ever-so-lame middle button
[19:32:41] <alex_joni> ok, whatever :)
[19:32:48] <alex_joni> I'm not sure we can ever agree
[19:32:57] <alex_joni> (not talking about myself personally)
[19:33:17] <alex_joni> it's most likely the code you talk about has been there for a while (meaning it works..)
[19:33:31] <maddash> if you look inside the gigantic switch() in emctaskissuecommand(), you'll see that unnecessary duplicates of the EMC_* classes are created
[19:33:40] <alex_joni> myself, I tend to leave old working code alone..
[19:33:58] <maddash> I said, "inefficient"
[19:33:58] <alex_joni> why do you think they are unnecessary?
[19:34:04] <maddash> e.g.,
[19:34:46] <maddash> for case EMC_AXIS_DISABLE_TYPE: "disable_msg = (EMC_AXIS_DISABLE *) cmd;
[19:34:46] <maddash> retval = emcAxisDisable(disable_msg->axis);"
[19:35:05] <alex_joni> btw.. the general consensous is that task should be completely replaced, but so far no-one attempted to do so (it contains a load of machine logic embedded in it)
[19:35:24] <alex_joni> madd so what's inefficient there?
[19:35:31] <maddash> then, if you follow the code, emcAxisdisable, when called, creates ANOTHER instance of EMC_AXIS_DISABLE class
[19:35:46] <maddash> stop being so antagonistic and follow the code
[19:36:01] <alex_joni> what do I get?
[19:36:17] <alex_joni> is there a cookie at the end of the path?
[19:36:20] <cradek> just object-oriented wankage
[19:36:31] <maddash> cradek: right.
[19:37:02] <alex_joni> maddash: ok, I followed and I saw there's a new instance there.. so what?
[19:37:02] <maddash> axis-disable's case statement and function call into *intf.cc isn't the only example here; these inefficient reinstantiations occur for ALL the cases of taskissuecommand's giant switch()
[19:37:21] <jepler> maddash: Which line is a "reinstantiation"?
[19:37:21] <alex_joni> did you look around to see what calls emcAxisDisable() ?
[19:37:28] <maddash> so why not pass on a pointter of the class itself?
[19:37:37] <alex_joni> jepler: it's hard to properly read
[19:37:47] <jepler> maddash: the line 'disable_msg = (EMC_AXIS_DISABLE *) cmd' just causes the pointer 'cmd' to be interpreted as pointing to existing memory which contains an EMC_AXIS_DISABLE structure
[19:37:47] <maddash> alex_joni: did you? try grepping
[19:37:51] <jepler> it does not allocate memory
[19:38:10] <alex_joni> maddash: and emcAxisDisable doesn't handle NML messages or classes
[19:38:16] <maddash> jepler: no, keep going into emcaxisdiable
[19:38:29] <alex_joni> it uses EMCMOT_DISABLE_AMPLIFIER
[19:38:47] <alex_joni> but maybe in your world EMCMOT_* is the same as EMC_*
[19:39:00] <alex_joni> of course looking at each definition might be too much work
[19:39:03] <maddash> bad example
[19:39:06] <xemet> I'm here
[19:39:07] <alex_joni> the EMCMOT_* are simple defines
[19:39:14] <alex_joni> the EMC_* are inherited classes
[19:39:16] <maddash> eg, case EMC_COOLANT_MIST_ON_TYPE
[19:39:43] <maddash> this is the case for all the cases which branch into iotaskintf.cc
[19:39:43] <alex_joni> maddash: there are a couple of things which send NML messages
[19:39:53] <alex_joni> stuff that goes to IO (like the coolant you mention)
[19:40:13] <xemet> jepler, did you succeed in tuning that: http://emergent.unpythonic.net/projects/01142347802
[19:40:24] <alex_joni> right.. but the code called is not always called because such a NML message was received
[19:40:35] <jepler> xemet: the tuning was never very good -- the use didn't require great following error
[19:40:38] <alex_joni> for example: coolant off is called on estop iirc
[19:41:01] <maddash> alex_joni: actually, it is. grep for emcCoolantMistOn inside src/emc/
[19:41:06] <xemet> ok, but you tune it and it woked
[19:41:16] <xemet> mine works *sometimes*
[19:41:42] <alex_joni> maddash: might be for mist on, but it's not for mist off
[19:41:48] <maddash> true, estop calls mistoff(), but nonetheless, these calls can be implemented better
[19:42:02] <alex_joni> and optimising 10% of all the code is worse than the current state
[19:42:14] <alex_joni> if that means you'll have an unregular API
[19:42:25] <alex_joni> some calls take numbers, others take pointers, etc
[19:42:30] <maddash> irregular*, and I said "these calls."
[19:42:35] <jepler> not that the current API is worthy of much love
[19:42:42] <alex_joni> jepler: I didn't say that
[19:43:01] <alex_joni> maddash: if you want to optimize, then figure out a way of reducing those pesky big case's
[19:43:19] <xemet> jepler, when your motor is stopped is it quiet or it vibrates and you hear like "trrrr"
[19:43:22] <maddash> really? I thought you LOVED the giant switch
[19:43:24] <maddash> es
[19:43:38] <alex_joni> jepler: but with the current API, once you start knowing it a bit, you can go on without many surprises
[19:43:44] <cradek> I think we would all love for someone to rewrite it
[19:43:52] <alex_joni> maddash: as I said, I *personally* would replace task completely
[19:43:53] <jepler> well unless he wrote something even nuttier
[19:44:06] <cradek> but, failing that, the next best thing is to leave the working code alone whenever possible :-)
[19:44:08] <alex_joni> but I wouldn't want to be the one doing that
[19:44:08] <maddash> ok, ok. so you three agree, which is enough for me.
[19:45:09] <maddash> brb, gotta stomp on the netadmin's foot
[19:46:02] <maddash> is there a reason why we're even using NML?
[19:46:24] <cradek> haha
[19:46:33] <maddash> maybe, a vestigial intention?
[19:46:45] <jepler> maddash: yes, it's the way the GUI communicates with task
[19:46:46] <anonimasu> no, the acronym were cool.
[19:46:47] <anonimasu> :D
[19:46:48] <jepler> perhaps you'd prefer to remove that part
[19:47:03] <jepler> I think that, after the Operator, the GUI is the single most error-prone part of emc
[19:47:04] <maddash> jepler: haha.
[19:47:16] <maddash> jepler: that was sarcasm, right?
[19:47:20] <jepler> which part?
[19:47:32] <maddash> 14:46:43 <jepler> maddash: yes, it's
[19:47:56] <alex_joni> no, that's the truth
[19:48:16] <alex_joni> most emc2 components communicate using NML
[19:48:28] <alex_joni> otoh, NML is used to do some other things aswell
[19:48:30] <maddash> and how is that relevant to my "why" question?
[19:48:50] <maddash> why not, straightforward shmem? dbus? FIFO?
[19:48:58] <alex_joni> on multiple PC's?
[19:49:06] <cradek> why do we have roads? that's where the cars drive
[19:49:29] <jmkasunich> EMC's original designers at NIST were working on distributed systems, and NML is key to that
[19:49:32] <maddash> cradek: wrong. we have roads, because those are the most efficient medium for car tires.
[19:49:33] <cradek> seems like a good answer to me
[19:49:46] <cradek> maddash: why do we have cars?
[19:49:48] <jmkasunich> just because most EMC users run everything on a single PC doesn't mean that everyone does it that way
[19:49:57] <maddash> cradek: why do we have a GUI?
[19:50:00] <cradek> cars are not efficient
[19:50:03] <alex_joni> maddash: this discussion has taken place at least a dozen times in the last 5 years
[19:50:17] <cradek> the answer is: that's what we had yesterday, so the easiest thing is to still have it today
[19:50:29] <maddash> it seems a bit cumbersome, which is why I'm asking.
[19:50:29] <jmkasunich> every new EMC developer hates NML and wants to replace it
[19:50:39] <alex_joni> maddash: I suggest you read through the logs if you want full answers.. you're not getting any from me (I'm tired of spelling it out)
[19:50:50] <jmkasunich> every new EMC developer realizes how much work that would be, and how many new bugs it would likely introduce
[19:50:59] <alex_joni> oh, and btw.. when I started on emc/emc2 I wanted to get rid of NML aswell
[19:51:00] <jepler> maddash: what cradek (and everyone) is trying to say is that there is a huge cost to getting rid of NML (e.g., you have a cnc system that doesn't work for at least a programmer-month, and doesn't work as well as it works now for a programmer-year)
[19:51:02] <jmkasunich> and every new EMC developer eventuall gets over if
[19:51:21] <maddash> especially when you have to wait for a response everytime you send off an NML request
[19:51:22] <alex_joni> jepler: you're talking full time
[19:51:24] <cradek> jepler is right
[19:51:38] <alex_joni> show me a programmer-year around OSS
[19:51:54] <anonimasu> maddash: well, it's nice to know your messages gets thróugh
[19:51:58] <cradek> and the payoff? some possible internal simplification to code that we don't even have to mess with much
[19:52:06] <anonimasu> maddash: what happens if you miss estop?
[19:52:12] <anonimasu> people die.
[19:52:19] <alex_joni> anonimasu: nope they don't :)
[19:52:30] <jmkasunich> anonimasu: don't go there - no sane person trusts _any_ software with estop
[19:52:30] <alex_joni> software is unreliable anyways.. so you shouldn't rely on this
[19:52:33] <maddash> anonimasu: there are better ways than, "Can you hear me now? ............................good."
[19:52:35] <jepler> (if the original emc hadn't been so inflexible to adapt to different types of machines, the same concerns would have killed emc2; but in that case, jmkasunich put in tons of work and essentially proved that his vision had a huge payoff)
[19:53:06] <anonimasu> jmkasunich, alex_joni: I think you guys missed the whole point
[19:53:14] <maddash> alex_joni: have a link to the logs?
[19:53:17] <alex_joni> maddash: what do you think would be the benefit
[19:53:21] <jepler> logger_emc: bookmark
[19:53:21] <jepler> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-12-28.txt
[19:53:26] <alex_joni> anonimasu: I didn't miss the point.. but estop is a lousy example
[19:53:53] <alex_joni> maybe if you would have said it misses a G0Z<safe> before G0x0y0
[19:53:59] <jmkasunich> right
[19:54:18] <maddash> jepler: how do I grep over tcp?
[19:54:21] <alex_joni> or a spindle on, before running into the part
[19:54:26] <xemet> jmkasunich, could you continua to help me or are you busy?
[19:54:35] <alex_joni> maddash: google search works great
[19:54:36] <jepler> maddash: I don't know; you'd better stop using the web though if that is important to you
[19:54:52] <alex_joni> search site:linuxcnc.org irc whatever
[19:55:04] <maddash> * maddash forgot about google. For once.
[19:55:08] <jmkasunich> give me 5 mins
[19:55:38] <anonimasu> jmkasunich: no, but there are plenty of insane people around
[19:55:57] <xemet> ok
[19:56:06] <maddash> the HAL pins don't use NML, do they?
[19:56:15] <alex_joni> maddash: nope
[19:56:32] <alex_joni> maddash: HAL is only .C programming ..
[19:56:56] <maddash> so it uses shmem
[19:57:19] <maddash> no reason why you can't extend the HAL API to C++
[19:57:32] <jmkasunich> maddash: its not a language thing
[19:57:34] <maddash> isn't HAL an IPC protocol? like NML?
[19:57:43] <alex_joni> maddash: one reason that comes to mind is that it's intended for a whole different purpose
[19:57:49] <jmkasunich> HAL is realtime, periodic updates, and signals flowing from one place to another
[19:57:52] <alex_joni> no, it's not
[19:58:01] <jmkasunich> NML is non-realtime, non-periodic, message oriented
[19:58:14] <tomp2> ?? from http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Halui "This Hal interface module provides the ability to connect HAL pins to NML commands."
[19:58:29] <jmkasunich> tomp2: right - it translates
[19:58:38] <jmkasunich> a change in the state of a HAL signal makes halui send a message
[19:58:41] <maddash> jmkasunich: i think the problem with nml comes from the 'message oriented' part
[19:58:53] <jmkasunich> or an NML message makes halui change the state of a hal signal
[20:00:04] <maddash> isn't that just like iocontrol.cc?
[20:00:17] <maddash> ugh, redundant
[20:00:26] <jmkasunich> not redundant
[20:00:35] <jmkasunich> yes, they both are hal <-> nml interfaces
[20:01:02] <jmkasunich> but iocontrol takes messages that are addressed to the "machine" and twiddles signals that go to the machine
[20:01:19] <jmkasunich> halui takes messages that are addressed to the operator and twiddles signals that go to the operator
[20:01:51] <maddash> why not have one module do both? instead of splitting it across two different threads?
[20:02:04] <jmkasunich> because EMC is modular
[20:02:06] <maddash> s/threads/modules/
[20:02:38] <jmkasunich> not everyone uses halui, some people do all operator interaction thru a gui, others want pendant buttons and such and use halui
[20:03:00] <jmkasunich> and why should iocontrol - machine logic stuff - be tied in with operator interface stuff?
[20:03:02] <alex_joni> others use halui on a different PC
[20:03:09] <maddash> too bad there's no way to knock off iocontrol...
[20:03:29] <alex_joni> maddash: I actually agree with you there..
[20:03:44] <alex_joni> it should be moved to motion (aka realtime)
[20:03:47] <jepler> it wouldn't be too terrible to see iocontrol go, replaced with hal pins exported by task or motion (not sure which one).
[20:03:48] <jmkasunich> we've thought about putting more (or all) of that stuff in motion
[20:03:49] <maddash> on this setup, all iocontrol does is take space and provide some estop loopback
[20:04:03] <maddash> motion
[20:04:05] <alex_joni> take space?
[20:04:06] <maddash> jmkasunich: motion
[20:04:07] <jmkasunich> maddash: "take space:
[20:04:08] <jmkasunich> ?
[20:04:08] <jepler> originally, iocontrol had to include all the machine-specific behaviors (for toolchange, etc)
[20:04:24] <jepler> but now this is all customized in hal instead of by changing and recompiling iocontrol
[20:04:25] <maddash> jmkasunich: pmem space. and cpu time.
[20:04:48] <jmkasunich> maddash: EMC is a PC based control
[20:04:56] <alex_joni> there were at least 3 different io controllers (spindle control, mist control, lube control, and regular control)
[20:05:12] <jmkasunich> in the PC, memory space and cpu time are (within reason) increasing faster than we are writing code
[20:05:31] <alex_joni> and emc2 has been started in order to make some of the code easier to read/follow, so we're not oriented at optimizing things too much
[20:05:54] <maddash> jmkasunich: only if you are rich enough to buy the newer boxes
[20:06:16] <jmkasunich> bull
[20:06:27] <alex_joni> I saw 5-600MHz boxes in dumpsters around here..
[20:06:28] <jmkasunich> I'm running my machine on a P3
[20:06:47] <alex_joni> and I'm not in the US..
[20:06:48] <maddash> alex_joni: 3 different io controllers? spindle/mist/lube are all inside iocontrol.cc...
[20:07:18] <alex_joni> maddash: do you read what I write?
[20:07:19] <jepler> hm, my 'io' process has a RSS of 1176 and has run for <.005% CPU over the last 3 hours, including startup time. That's pretty light on resources by today's standards--1/3 the rss of my shell, for instance.
[20:07:33] <maddash> alex_joni: no, xrandr fscked my dpi
[20:07:46] <alex_joni> it used to have (in emc1)
[20:08:19] <alex_joni> all IO controllers used to talk through different NML channels.. that was a waste imo
[20:09:03] <maddash> yeesh.
[20:09:11] <alex_joni> but even that.. wasn't that noticeable from a resource point of view
[20:09:20] <alex_joni> (and that was 3-4 years ago)
[20:09:30] <jmkasunich> try 10+ years ago
[20:09:38] <jmkasunich> emc1 dates from 1995 or earlier
[20:09:54] <alex_joni> I didn't use it 10+ years ago.. so I don't know if it was a resource hogger back then :)
[20:10:13] <alex_joni> but 3-4 years ago when I first tried it.. it wasn't :)
[20:10:27] <alex_joni> http://www.youtube.com/watch?v=AoOWi1r6-Eo <- yay#2 emc2
[20:10:45] <jmkasunich> xemet: can you do private messages? (I think you have to be registered on freenode)
[20:11:03] <maddash> i shall now be silent.
[20:11:12] <xemet> I've replied
[20:11:17] <maddash> ...and work on emctaskmain.cc
[20:12:05] <alex_joni> maddash: still.. if you find a solution that works good enough in replacing NML (keeping all the current features), then I see no reason against that
[20:12:33] <anonimasu> hehe
[20:12:58] <Dallur> alex_joni: way cool, what are you using that arm for ?
[20:13:24] <alex_joni> Dallur: not mine..
[20:14:11] <Dallur> can you tell us what it's intended use is though ?
[20:14:14] <alex_joni> a guy from germany built it
[20:14:19] <alex_joni> I think it's just for fun :)
[20:14:33] <alex_joni> it took about a year to get to this point
[20:14:40] <alex_joni> maybe more
[20:14:52] <alex_joni> (including mechanical work, work on emc2, kins, etc)
[20:15:09] <maddash> oooh, it can play Hanoi
[20:15:36] <alex_joni> I guess hanoi would be quite easy using o-words
[20:15:39] <Dallur> Who's hooking a chess program up to EMC2 so that arm can play chess ?
[20:15:51] <jepler> shttp://pastebin.ca/835296
[20:15:56] <jepler> er, s/^s//
[20:16:12] <alex_joni> 6-level hanoi and it'll be occupied for a while
[20:16:30] <cradek> haha gnome-terminal
[20:17:08] <maddash> rofl gnome
[20:17:34] <Dallur> rofl kde :)
[20:17:41] <alex_joni> jepler: is VSZ in byte?
[20:17:47] <alex_joni> or kByte?
[20:17:50] <jepler> alex_joni: kilobytes I think
[20:18:21] <alex_joni> hmm.. 3MB does feel like much for io
[20:18:49] <maddash> milltask takes up 2-9% cpu
[20:19:51] <maddash> alex_joni: this is ps, so it's ignoring the fact that there's some library sharing going on
[20:20:06] <alex_joni> I bet at least half of that is interp.. (if running)
[20:22:36] <maddash> ugh, gnome. vile creation.
[20:23:17] <maddash> here we go
[20:23:20] <cyborg_ar> hello
[20:23:44] <maddash> AXIS not playing your mp3s again?
[20:23:55] <maddash> j/k.
[20:24:23] <cyborg_ar> mmm i use ogg vorbis
[20:25:12] <maddash> anyone read italian? http://www.emc2cnc.altervista.org/
[20:25:15] <alex_joni> well.. it does play mine.. actually it engraves them
[20:25:31] <jepler> actually you should use halstreamer to copy your pcm data into kernel space, use a custom component to do 1-bit sampling (sigma-delta or whatever it's called) and connect that to hal_speaker -- voila, high-quality audio that you can listen to while you use emc.
[20:25:44] <maddash> omg, that's xemet's page
[20:25:54] <alex_joni> a simple filter that opens mp3's runs them through fft, and creates XY engravings
[20:26:04] <maddash> xemet: you're my hero. how'd you get the NURBs working?
[20:26:14] <cyborg_ar> an how can i do rip cds whit axis?
[20:26:21] <alex_joni> maddash: more coding than talking :)
[20:26:25] <cyborg_ar> a lasser scanner
[20:26:26] <cyborg_ar> ?
[20:26:39] <alex_joni> cyborg_ar: place CD on table, use a very fine touch-probe
[20:26:54] <cyborg_ar> mmm that works for lp
[20:26:56] <alex_joni> heat it up so it goes through the plastic..
[20:27:03] <maddash> alex_joni: true. I was hoping that he'dfound some other way besides brute-force interpolation and libposemath
[20:27:12] <alex_joni> and when you hit the metal part of the cd you know your position :)
[20:27:15] <lerneaen_hydra> alex_joni, now that would be sweet, a filter that goes from mp3/ogg/whatever and generates a vinyl-like engraving :D (with options for track width, pitch, amplitude and so on)
[20:27:35] <cyborg_ar> cool
[20:27:37] <alex_joni> lerneaen_hydra: using a rotary axis or not?
[20:28:07] <lerneaen_hydra> alex_joni, accuracy would be better with rotary, though it may be doable with a normal 3-axis machine
[20:28:08] <alex_joni> quite a lot of videos here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Videos
[20:28:10] <alex_joni> bbl..
[20:28:23] <lerneaen_hydra> with an insanely large pitch and amplitude
[20:28:35] <lerneaen_hydra> so the system's resolution is reduced
[20:28:39] <cyborg_ar> hehe
[20:28:48] <lerneaen_hydra> err
[20:28:59] <lerneaen_hydra> so the effect of the systems finite resolution is reduced
[20:29:41] <lerneaen_hydra> ie. sounds like shit with some music superimposed in contrast to just noise
[20:29:48] <cyborg_ar> how about usin axis and a laser plotter for burning cd's?
[20:30:40] <lerneaen_hydra> that's cheating, you have to use a V tipped endmill
[20:30:48] <lerneaen_hydra> and physically engrave it
[20:31:41] <cyborg_ar> oh, and then put the transparent plastic over the engraved metal
[20:31:54] <lerneaen_hydra> oh I was thinking vinyl now
[20:32:22] <lerneaen_hydra> or any other material, just as long as you have a mechanical pickup arm as vinyl disks do
[20:39:46] <Dallur> is any of the nurbs code in cvs yet ?
[20:40:17] <jepler> Dallur: no
[20:40:56] <BigJohnT> $2.50 gecko heat sink with fan! http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/HPIM0476.jpg
[20:40:58] <Dallur> jepler: Any idea when/if Manfredi intends to submit ?
[20:41:39] <Dallur> jepler: to bad he does not seem to be at his computer :) well, might catch him later
[20:46:30] <jepler> Dallur: I don't know his plans. I haven't talked to him about it in quite awhile, but I think the code still has the limitation that the subdivision is arbitrary, not adaptive; is only in the XY plane; and does not support linear motion in the third axis ("helical bezier curve"). I'd like all three of those things to be fixed before the code is included.
[20:47:17] <jepler> er, I guess "helical nurbs curve" since he used a more general curve type than I used in my original code
[20:47:27] <Dallur> jepler: yeah, it's also all non-realtime, looks like it's converting the splines into curves non-adaptively
[20:48:07] <cradek> even if he doesn't want to do all those things right now, or ever, putting the code in cvs, even on a branch, would be very nice
[20:48:08] <Dallur> http://www.cncitalia.net/forum/viewtopic.php?t=1179&sid=04c081872f34ad68dcd2072d3e07e8f5
[20:48:08] <maddash> RT nurbs interpolation is pointless
[20:48:22] <cradek> someone else may like to finish that stuff
[20:48:39] <maddash> if you really want to cut along a spline, why not approximate with biarcs and use those instead?
[20:49:02] <Dallur> maddash: true, but I think it should be "just in time" and try to use the resolution as calculated from your emc parameters to create the trajectory
[20:51:14] <Dallur> maddash: IMHO if it's done earlier and approximate g-code generated then it can just as-well be done during the CAD/CAM phase
[20:51:52] <maddash> Dallur: I was referring to the CAM phase in my remark
[20:55:57] <BigJohnT> Dallur: What does the float switch do in your plasma torch?
[20:56:05] <Dallur> maddash: I see, my primary interest in the "Just in Time" concept is the portability and scalability of the g-code, if you can naturally represent splines it makes it possible to get the best possible resolution on any machine and scale the size of the piece when working in a 2/2.5D environment
[20:57:27] <Dallur> BigJohnT: The float switch is the switch that tells the machine that the torch is X distance away from the surface of the work piece, it's kinda like a probe switch except in most cases the switch never touches the work piece but would be a weld safe inductive or capacitive sensor
[20:58:12] <BigJohnT> Dallur: Ok that makes sense thanks
[20:58:26] <Dallur> BigJohnT: np
[20:59:23] <BigJohnT> Dallur: this huge print of your NewPlasmaDesign is lot's easier to read LOL
[21:01:04] <Dallur> BigJohnT: I know it's all a bit overwhelming but it's the best way I found to document the hal components
[21:02:35] <BigJohnT> Dallur: it is very good and easy to follow now that it's big enough for me to read... I copied your style for my plasma torch and started drawing it up as you did makes it easier to follow. Thanks
[21:45:22] <rcsu> hi *
[21:45:36] <rcsu> * rcsu is just a bit lurking
[21:46:03] <jepler> lurk all you want
[21:56:04] <gezr> im are classes registered
[21:56:35] <gezr> 18 hours at that
[21:57:54] <gezr> anything fancy happening today?
[21:58:31] <BigJohnT> $2.50 heat sink and fan for Gecko's
[21:58:42] <gezr> thats all?
[21:59:24] <BigJohnT> wanna see?
[21:59:27] <gezr> and I think we gamma-x going in the right direction on his mill last night,
[21:59:35] <gezr> yes I would like to see it BigJohnT
[21:59:49] <BigJohnT> good
[22:00:07] <BigJohnT> took a while to load the page at home in the woods http://i47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/HPIM0476.jpg
[22:00:35] <gezr> thats pretty sweet
[22:00:49] <gezr> you gonna have like a drop down mount for it?
[22:00:51] <alex_joni> gezr: I put some nice vids on youtube
[22:00:52] <BigJohnT> know what it is
[22:01:02] <gezr> cpu sink and fan :)
[22:01:19] <BigJohnT> yep been machined on some
[22:01:21] <gezr> alex? oh link
[22:01:24] <BigJohnT> socket a
[22:01:27] <alex_joni> youtube.com
[22:01:30] <gezr> bigjohn thats not a bad idea
[22:01:31] <alex_joni> gezr: :P
[22:01:35] <BigJohnT> I like the robot one alex
[22:01:45] <alex_joni> http://www.youtube.com/profile?user=alexjoni
[22:01:46] <gezr> alex under your name alex_joni?
[22:04:03] <gezr> thats freaken sweet
[22:05:14] <gezr> oh I got all my data moved onto 1 drive last night, got it all down to about 1.3g in size, then blew out some stuff on the old linux install, got me 20g drive to use as an emc/linux drive now :), im really excited
[22:05:33] <gezr> does the ubuntu 2.2.2 live disks install come with dev tools?
[22:05:44] <alex_joni> not on the cd
[22:05:49] <alex_joni> but you can apt-get them later
[22:06:04] <gezr> okay, anything In particular I should look out for?
[22:06:15] <alex_joni> apt-get build-dep emc2
[22:06:40] <gezr> ill probably start the install in about an hour or so
[22:06:43] <alex_joni> (that gives you all packages required to build the emc2 package yourself)
[22:07:05] <alex_joni> if you don't want to do that.. then you can simply use the binaries
[22:07:13] <gezr> I love compiling
[22:07:27] <gezr> and figuring stuff out, besides, im going to want to piddle with the code
[22:17:24] <alex_joni> this is a nice router: http://www.wonti.de/werkstattFraese_CNC.html
[22:19:35] <alex_joni> and he's using it to build some really nice toys: http://www.wonti.de/mod_radlader_start.html
[22:27:36] <BigJohn1> I just plugged a couple of parport cards in and disabled the onboard parport. How do I tell if the cards are seen by linux?
[22:28:22] <alex_joni> BigJohn1: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?NetMos
[22:28:52] <BigJohn1> alex_joni: thanks
[22:49:02] <alex_joni> good night all
[22:53:19] <maddash> a small correction to line 819 of http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/emc/iotask/ioControl.cc?annotate=1.45
[22:53:41] <maddash> jepler/cradek ^^
[22:54:06] <maddash> it's trivial, but the semi-colon shouldn't be there.
[23:03:46] <gezr> i have begun the install
[23:05:06] <gezr> there we go, i/o errors on the cdrom sigh
[23:05:27] <BigJohnT> I'll trade you for parport problems
[23:06:40] <gezr> what is device dm-1?
[23:07:21] <gezr> well it made it past the errors
[23:07:32] <gezr> big what sort of paraport problem are you having?
[23:07:47] <gezr> it could be a simple bios adjustment being needed, have you looked at that?
[23:08:15] <BigJohnT> well the out 7,8,9,14,16 don't seem to work on the onboard parport and ...
[23:08:32] <BigJohnT> I get the same results from a plug in parport
[23:08:52] <BigJohnT> yea, played with all the bios settings
[23:09:23] <BigJohnT> something fishy going on and I don't know what it is
[23:10:03] <BigJohnT> gezr: you ever use a thumb drive with linux?
[23:10:14] <gezr> no i havent
[23:10:31] <BigJohnT> I plugged mine in but could not find it with the file browser
[23:13:39] <gezr> youll have to mount it first
[23:14:25] <gezr> and I dont know how to do that with usb devices
[23:15:21] <gezr> mkdir /mnt/thumbdrive then mount /dev/usb1 /mnt/thumbdrive but thats probably no where near correct
[23:16:30] <gezr> ive never really ran linux on a machine with usb
[23:16:42] <gezr> im about to, but still, I dont have a thumbdrive
[23:17:22] <BigJohnT> thanks
[23:18:25] <BigJohnT> anyone, Ok I ran sudo modprobe -a parport_pc to see what the base address is of my pci parallel port card is so where is the "kernel message log"
[23:20:59] <BigJohnT> gezr: you trying to do an install?
[23:22:40] <gezr> yeah
[23:22:56] <gezr> cat /var/log/messages
[23:23:29] <BigJohnT> holy crap!
[23:23:33] <gezr> ?
[23:24:39] <BigJohnT> got pages of stuff
[23:24:58] <gezr> set this in a new terminal window tail -f /var/log/messages
[23:25:17] <gezr> then in the other window you can run your other commnands and as messages gets written to, you will see the output
[23:25:33] <gezr> to end the tail just press ctrl-c
[23:27:23] <BigJohnT> cool
[23:29:18] <gezr> enterprise volume management system, then I get i/o erros on device dm-2...
[23:29:22] <gezr> what in the hell is that
[23:30:49] <BigJohnT> dunno
[23:31:50] <gezr> ah something with raid arrays
[23:32:31] <gezr> alrighty, another problem, root password?
[23:32:46] <BigJohnT> you gotta love it
[23:33:16] <gezr> so they disabled the root user?
[23:34:09] <gezr> this is not going to make me happy one bit
[23:34:42] <gezr> bam root
[23:34:52] <gezr> now I can root around happy town
[23:35:57] <BigJohnT> I got a .357 you can borrow
[23:36:44] <gezr> oh i fixed the whole root issue
[23:36:48] <gezr> heh
[23:42:42] <ejholmgren> heh
[23:42:58] <ejholmgren> dirty harry that thing
[23:45:46] <gezr> well the sim axis works just fine
[23:46:09] <gezr> ill have to piddle with the software for a while, but im going to let it simmer on the burner for a bit