#emc | Logs for 2007-02-23

[00:01:23] <twice2> jepler: i guess my question is in what units is MAX_VELOCITY?
[00:01:39] <jepler> units (inifile units) per second
[00:02:24] <jepler> generally inches or mm
[00:02:54] <twice2> ok, in my case input_scale is steps/inch
[00:03:41] <twice2> so 28800 steps per inch times 0.30 inches per second give me........
[00:04:07] <jepler> 8640 steps per second
[00:04:19] <jepler> possible with even a modest BASE_PERIOD
[00:04:19] <twice2> not what emc says
[00:04:56] <jmkasunich> what does EMC say?
[00:05:08] <jmkasunich> (exact wording please, if possible copy and paste the message)
[00:05:09] <twice2> it's telling me that i requested 42040~ steps per second
[00:06:08] <jmkasunich> can you put your config files at pastebin.ca?
[00:06:28] <jmkasunich> http://pastebin.ca/
[00:06:43] <jmkasunich> copy the files, paste them into the window, submit, then paste the URL here
[00:07:17] <twice2> i could if my german shepherd will leave me
[00:08:51] <jepler> I changed the standard stepper config to have the MAX_VELOCITY and INPUT_SCALE you said, and to have a STEPGEN_MAXVEL of 0.33 (10% headroom). http://pastebin.ca/368331
[00:08:53] <twice2> ok, i'm going to post my stepper.ini it's the only one that i've change so far
[00:09:31] <jepler> even with the BASE_PERIOD unchanged (50000ns) it does not give the error
[00:10:13] <twice2> ok, good stuff, let me post mine, it would be great if you would look at it
[00:10:31] <jepler> If I increase STEPGEN_MAXVEL to 1.4 (I believe this may be the value in the sample config) I get this error:
[00:10:33] <jepler> STEPGEN: Channel 0: The requested maximum velocity of 40319 steps per second is not attainable.
[00:10:36] <jepler> STEPGEN: The maximum number of steps possible is 10000 per second
[00:10:46] <jepler> 17:51:35 <jepler> the emc 2.1 "requested velocity .. is not attainable" message is based on STEPGEN_MAXVEL, not MAX_VELOCITY
[00:12:22] <twice2> http://linux.pastebin.ca/368332
[00:15:45] <jmkasunich> MAX_VELOCITY = 0.243, STEPGEN_MAXVEL = 1.294
[00:16:07] <jmkasunich> set STEPGEN_MAXVEL to about 10% more than MAX_VELOCITY, try 0.27 or so
[00:18:31] <twice2> oh, with that ini it runs fine. i'm trying to top out ipm
[00:19:09] <jmkasunich> ?
[00:19:09] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TweakingSoftwareStepGeneration
[00:19:24] <twice2> where do i read this, it was accel
[00:19:34] <jmkasunich> what?
[00:19:38] <SWPadnos> you need to start by figuring out what your motors/drivers can do, and what your PC can do
[00:20:03] <SWPadnos> then set the ini parameters to match that (or lower, if the attainable speed is beyond what you want)
[00:21:17] <twice2> yeah, that's it :) I had them running 22 ipm but would stall on some radips
[00:22:30] <jmkasunich> you either have to reduce speed, or reduce accel (or both)
[00:22:31] <twice2> ok cool beans, thanks for the link there
[00:22:46] <SWPadnos> sure. and thanks to jmkasunich for writing the page ...
[00:22:48] <jmkasunich> to figure out which, you need to take one of the out of the picture
[00:23:03] <jepler> yeah -- I hadn't read that page before, but like all of his work it is very thorough
[00:23:12] <jmkasunich> you could try setting accel quite low, and experimenting with velocity first
[00:23:22] <jmkasunich> once you find a velocity that it can do, then increase the accel
[00:24:10] <twice2> i've seen both your nics everywhere, thanks for taking the time to develop emc, it's cool stuff
[00:24:22] <jmkasunich> you're welcome
[00:29:33] <ds3> does emc support using lines off the serial port?
[00:31:08] <jmkasunich> you mean as general purpose I/O bits? or as serial communications?
[00:31:16] <jmkasunich> general purpose I/O, yes
[00:31:48] <ds3> GP IO
[00:32:08] <jmkasunich> serport is the driver for that
[00:32:27] <ds3> any obvious restrictions compared to the parallel port?
[00:32:27] <jmkasunich> man serport should tell you a little about it
[00:32:31] <ds3> gotcha
[00:32:42] <ds3> serial ports are more robust ;)
[01:14:37] <jepler> ds3: "man serport" in emc2.1
[01:15:10] <jepler> or http://linuxcnc.org/docs/2.1/html/man/man9/serport.9.html
[01:35:49] <Ziegler> hey guys
[01:36:02] <Ziegler> looking to edit arch tolerance
[01:36:15] <Ziegler> is that something I can change in my INI file?
[01:36:22] <Ziegler> err arc*
[01:36:55] <crepincdotcom> in the tradjectory planner
[01:37:00] <crepincdotcom> as ERROR, as i recall
[01:37:06] <crepincdotcom> look through the options, its prety obvious
[01:38:10] <Ziegler> *looking again*
[01:39:03] <crepincdotcom> at least i recall it being so, it's been a bit since i messed with it
[01:42:10] <Ziegler> hmm... not seeing anything
[01:42:39] <Ziegler> maybe try ERROR = .001 ?
[01:42:39] <crepincdotcom> I'm not at my emc box atm, sorry can't check for you
[01:43:02] <crepincdotcom> or look in each axis' config section, is there a MIN_ERROR and MAX_ERROR?
[01:44:31] <Ziegler> FERROR = 0.050
[01:44:32] <Ziegler> MIN_FERROR = 0.010
[01:44:35] <Ziegler> is that it?
[01:44:44] <skunkworks> jepler: are you around?
[01:45:00] <crepincdotcom> Ziegler: beleive so yes
[01:45:03] <crepincdotcom> see the docs about those
[01:45:17] <Ziegler> ok thank you.. I wil lgive that a search
[01:45:28] <crepincdotcom> sure
[01:49:03] <jepler> Ziegler: specify a tolerance number to G64 P-
[01:49:15] <Ziegler> hmm.. tried that too.
[01:49:50] <Ziegler> g64 p-0.001
[01:50:08] <jepler> specify a positive number, the "P-" notation means "P and a number".
[01:50:34] <jepler> I am not sure what specifying a negative number actually does
[01:51:03] <jepler> skunkworks: I'm around now
[01:51:08] <crepincdotcom> * crepincdotcom votes segfault
[01:51:14] <Ziegler> ok let me play with it a bit
[01:53:50] <Ziegler> I even tried a G64 P1 and I still get a "Radius to end of arc differes from radius to start"
[01:54:08] <crepincdotcom> are you sure your endpoints are correct?
[01:54:23] <jepler> oh -- G64 P- won't fix that error
[01:54:29] <jepler> neither will FERROR and MIN_FERROR
[01:54:37] <Ziegler> If I cange to G21 instead of G20.. it will work
[01:54:38] <crepincdotcom> like does that happen if you do"g0 x1 y1 [next line] g2 x1 y1 i-1 y0"
[01:54:38] <Ziegler> but its in mm then
[01:55:14] <Ziegler> G01 Z-0.2
[01:55:18] <Ziegler> G03 X2.413 Y0.414 I0.197 J0.610
[01:55:19] <jepler> "When the arc is projected on the selected plane, the distance from the current point to the center differs from the distance from the end point to the center by more than 0.0002 inch (if inches are being used) or 0.002 millimeter (if millimeters are being used)."
[01:55:20] <Ziegler> G03 X2.721 Y1.468 I-1.279 J0.945
[01:55:30] <Ziegler> seems to happen at teh second line
[01:55:50] <jepler> "Use at least 3 digits after the decimal when milling in millimeters, and at least 4 digits after the decimal when milling in inches. In particular, arc tolerance checks are made to .001 and .0001 depending on the active units. "
[01:55:54] <crepincdotcom> * crepincdotcom votes code issue
[01:56:07] <skunkworks> jepler: should halmeter work to view position from the pluto? it is doing odd things.
[01:56:08] <jepler> (http://linuxcnc.org/docs/2.1/html/gcode/main/)
[01:56:11] <crepincdotcom> try the two lines i said
[01:57:25] <Ziegler> yeah I came across that earlier tonight... is there anyway to adjust the program acceptable toleance since my g-code is all at 3 decimal?
[01:57:52] <jepler> Ziegler: not without recompiling.
[01:58:00] <Ziegler> lol
[01:58:06] <crepincdotcom> awk and sed are your friends
[01:58:17] <jepler> crepincdotcom: I do not believe they will help
[01:58:31] <jepler> simply tacking on an extra zero of precision, for instance, will not change the result.
[01:58:41] <crepincdotcom> ahh i see now
[01:58:44] <crepincdotcom> never mind then
[01:58:52] <Ziegler> (yeah.. attempted it already ;-) )
[02:16:45] <skunkworks> jepler: ? when I look at position in halmeter - it jumps around between 2 different numbers. what seems like the encoder count and some other lower count.
[02:17:58] <skunkworks> if I turn it one rotation I get about 4000 which is right - then 8000 for 2 and then it looses it (numbers don't make sense.) I can usually turn it back to reading about 0 - so it doesn't seem to be loosing it - it just doesn't make sense
[02:18:46] <jepler> skunkworks: hm -- that's not good
[02:19:15] <jepler> skunkworks: does it "jump" when the encoder is stopped?
[02:19:21] <jepler> skunkworks: or only while it's turning?
[02:19:29] <skunkworks> all the time.
[02:19:40] <skunkworks> stopped or moving.
[02:19:44] <jepler> can you find the two numbers and tell me the difference between them?
[02:20:20] <skunkworks> right now it it flashing between -64 and 76
[02:20:33] <jepler> is one of the numbers always negative?
[02:20:45] <skunkworks> now after I turn it a bit it is flashing between -202 and 1
[02:20:50] <skunkworks> let me see
[02:23:07] <skunkworks> now it is flashing between 64 and 213
[02:23:46] <skunkworks> at about 1 rotation it is flashing between 4016 and 79
[02:24:49] <jepler> how easy would it be to test on the encoder 2 input (QA2/QB2) instead?
[02:25:26] <skunkworks> easy - but I will be back in a bit.
[02:25:47] <jepler> crap, I hope the pluto isn't completely broken in 2.1.0
[02:35:20] <cradek> why do you say that?
[02:35:35] <jepler> cradek: skunkworks is getting some weird behavior from it
[02:35:59] <cradek> why would you think it isn't just skunkworks that's the problem??
[02:36:13] <jepler> cradek: uh, chris, he's back already, you can't say that
[02:36:16] <cradek> * cradek pokes skunkworks
[02:36:18] <cradek> haha
[02:36:34] <cradek> want me to go try mine?
[02:36:43] <jepler> if you're not busy
[02:37:09] <cradek> the encoder counting is the question?
[02:37:13] <jepler> yes, I think so
[02:37:24] <cradek> not sure if I'm set up, but I'll go see
[02:38:06] <skunkworks> sorry/
[02:38:26] <skunkworks> I tried the a1/b1 and get similar results/
[02:38:53] <jepler> skunkworks: it's a2/b2 I want you to try, because the "debug" parameters show extra info about that encoder
[02:39:25] <jmkasunich> hey - what is a simple, user friendly IRC client for doze?
[02:39:28] <skunkworks> 0,1,2<- is what you want
[02:40:02] <cradek> jmkasunich: I think mirc is what everyone uses - it looks like xchat
[02:40:55] <jepler> skunkworks: with the encoder hooked up to qa2/qb2, look at the "debug" params once you're in the freak-out state. If the "debug" numbers are changing then it's more likely to be a firmware issue; if they're not, it's more likely to be a driver issue.
[02:42:17] <jepler> the firmware's internal quadrature count (which can only count up to 8192 before something "special" happens in the driver code) is displayed on the pin debug-0
[02:42:25] <cradek> HAL:7: ERROR: Can't find module 'pluto_servo' in /usr/realtime-2.6.15-magma/modules/emc2
[02:42:29] <cradek> uh?
[02:42:31] <cradek> HAL config file /home/chris/emc2/configs/lathe-pluto/lathe-pluto.hal failed.
[02:42:35] <skunkworks> ok - remember who your talking to - I am just using the pluto.hal - that make the led flash
[02:42:57] <skunkworks> can I get the the debug mode though here?
[02:43:10] <jepler> skunkworks: yes, this is a HAL pin I'm talking about. "pluto-servo.debug-0" I think
[02:43:32] <jepler> or parameter
[02:43:35] <jepler> HAL parameter
[02:43:51] <jmkasunich> cradek: what tree? head? 2.1? I have it in both
[02:43:59] <cradek> installed 2.1.0
[02:44:05] <cradek> it's not there, and it's not in the deb
[02:44:07] <jmkasunich> oh, its looking in /usr?
[02:44:22] <jmkasunich> sorry, I was looking at my cvs checkout
[02:44:25] <skunkworks> it does the same thing - flashes between 2 different numbers - Is it something goofy with my setup?
[02:44:47] <cradek> I'll build a rip - may be a while
[02:44:58] <cradek> but it sure isn't in the package
[02:45:08] <jmkasunich> another item for 2.1.1
[02:45:38] <jmkasunich> do things explicitly need to be added to some control file to make them appear in the deb?
[02:45:47] <jepler> skunkworks: OK, thanks
[02:45:54] <jepler> maybe I'll be able to spot the problem based on that
[02:46:23] <jepler> jmkasunich: yes, it should be a line in emc2.files.in but it isn't
[02:46:30] <cradek> drivers.files.in?
[02:46:43] <jepler> or that
[02:46:48] <skunkworks> The debug-0 looks like similar to the 3 encoder position. 1 rotation flashes between 4097 and 60
[02:47:19] <jlmjvm> has anyone here used a pci express video card with emc 2.1 yet?
[02:47:34] <jepler> not me -- I only have one machine that new and it's not for emc
[02:47:42] <cradek> jlmjvm: nope
[02:47:44] <skunkworks> sometimes it just looses it and goes to say 19000 or more and I can never get it back to aprox 0
[02:48:20] <cradek> jlmjvm: plenty of regular PCI though - you might try that if you have one. seems the simpler/older video cards are better
[02:48:51] <jlmjvm> which gpu do you think would be better,nvidia,or ati
[02:49:06] <jepler> skunkworks: oh drat -- the debug information isn't showing me what I thought it was
[02:49:11] <jepler> so it's not that useful
[02:49:17] <cradek> you may have to use the VESA driver with either one - neither will be much better than the other
[02:49:37] <cradek> seems like an old PCI matrox card is a good bet - a lot of us use those and they are everywhere
[02:49:57] <jlmjvm> k,will keep that in mind,will look for 1
[02:50:16] <skunkworks> the pwm0 out look correct in halmeter.
[02:50:20] <cradek> I should bring a stack of Millenium IIs and a stack of external modems to give away as door prizes at fest
[02:50:35] <jepler> skunkworks: can you make a 1-line change to pluto_servo.comp and recompile? http://pastebin.ca/368470
[02:50:37] <jmkasunich> jepler: the extend() function seems a bit wonky to me
[02:50:40] <skunkworks> sweeping through -1 and 1
[02:51:02] <jmkasunich> can't say its not correct, but it isn't what I would expect for sign extension
[02:51:48] <jmkasunich> not sign extension... you know what I mean
[02:52:01] <jepler> jmkasunich: I didn't find a "right way" when I looked so I wrote what I came up with after doodling on paper
[02:52:16] <jmkasunich> * jmkasunich doodles too
[02:52:18] <jepler> jmkasunich: my little test program didn't find a problem but it probably had the same blind spot I did
[02:52:33] <jmkasunich> the counter is 14 bits?
[02:53:10] <jepler> jmkasunich: yes
[02:53:17] <cradek> jepler: it seems totally broken here too
[02:53:31] <jepler> cradek: oh good
[02:54:16] <cradek> should I try trunk?
[02:54:25] <jepler> I don't think it's any different
[02:54:39] <skunkworks> sorry - where is pluto.com?
[02:54:40] <skunkworks> comp
[02:54:47] <jepler> skunkworks: src/hal/drivers/pluto_servo.comp
[02:55:20] <cradek> like an idiot I just ran cvs up in my working tree...
[02:55:25] <jlmjvm> there are actually quite a few pci cards on newegg
[02:55:51] <cradek> (at least I think it was working)
[02:56:58] <jmkasunich> jepler: I think my preferred way is: longval += newshort - oldshort
[02:57:10] <jmkasunich> but that requires storing oldshort
[02:57:12] <jepler> it looks like the extend() function hasn't changed since the oldest revision checked in
[02:57:54] <jmkasunich> extend could be perfectly fine - there are many ways to skin that cat
[02:58:11] <jmkasunich> I just thought it interesting that it gets wonky at 4096 and 8192 or in that neighborhood
[02:58:24] <jepler> yes I think it's interesting too
[02:58:46] <skunkworks> jepler - one more question - how do I recompile
[02:59:00] <jmkasunich> hack the driver so the value read from the hardware is on a parameter
[02:59:12] <jepler> skunkworks: type "make" in the src/ directory
[02:59:28] <jmkasunich> use sim encoder to make quadrature and send it out the parport, loop it back to pluto, and scope it
[02:59:33] <jmkasunich> pain in the ass, I know
[02:59:41] <jepler> jmkasunich: there are digital outs on the pluto that I think I trust
[02:59:45] <jmkasunich> actually, you could send the sim encoder out the digital IO on the pluto
[02:59:47] <jmkasunich> ;-)
[03:00:52] <cradek> here, encoder-0-count shows 0 1 or 2; the others don't do anything
[03:01:13] <jmkasunich> I guess extra semicolons in C don't hurt anything:
[03:01:14] <cradek> on both trunk and RELEASE_2_1_0
[03:01:15] <jmkasunich> count = extend(data.last_count[0], ppdata & 0x3fff);;
[03:01:39] <skunkworks> ok - what do you want me to do? It looks like it is showing the same info
[03:02:17] <jepler> skunkworks: they still read the same after "freak out"?
[03:03:05] <jepler> are you using run-in-place or run-installed?
[03:03:12] <jepler> if the latter, you need to "make install" after "make"
[03:03:17] <skunkworks> run in place
[03:03:26] <skunkworks> oops
[03:03:47] <skunkworks> oh - I am fine - run in place
[03:04:16] <skunkworks> it is always freaking out in my opinion - it is constantly flipping between 2 different numbers.
[03:05:14] <skunkworks> always is maybe not correct. it flips between 2 different number most of the time. all the time when it is rotating and most of the time while it is stopped
[03:05:17] <jepler> oh -- when you first explained the trouble you were having, I thought you meant it behaved normally until you got to 8000 (two revs)
[03:06:05] <skunkworks> sorry - no it flips between 2 numbers from the get-go
[03:06:49] <SWPadnos> can you turn the shaft slowly enough to see the prograssion of the numbers?
[03:06:55] <SWPadnos> progression
[03:07:00] <skunkworks> one of the numbers seems to be position. -
[03:08:23] <skunkworks> I can turn it one rotation and get aprox 4000 - then 2 rotations and one of the numbers is 8000 - and then I can turn it back to aprox 0 - all the time there is this other number that it alternatlly displays that doesn't make sense to me.
[03:08:34] <Jymmmm> For the record, the "Interlocking Connectors" at Radio Shack are 0.093" diameter Molex, Not Tyco/Amp.
[03:09:09] <SWPadnos> skunkworks, ok, but can you start it at 0 and tell us what the sequence is?
[03:09:22] <Jymmmm> greycode
[03:09:35] <SWPadnos> wouldn't that be funnt
[03:09:40] <SWPadnos> s/t/y/
[03:09:54] <Jymmmm> I wouldn't be surprised, kinda makes sense.
[03:10:07] <SWPadnos> not for a binary up/down counter
[03:10:20] <Jymmmm> binary no, rotary, yes.
[03:10:28] <skunkworks> ok this is encoder posision
[03:10:33] <skunkworks> position
[03:10:37] <cradek> I have a *working* checkout from 2006-12-14
[03:10:40] <SWPadnos> it's the encoder counter that we're checking :)
[03:11:18] <jmkasunich> skunkworks: does the other number change?
[03:11:19] <skunkworks> just started - 0 (no flashing)
[03:11:58] <skunkworks> turned it clockwise 1/4 turn. flashing between 1036 and 4
[03:12:09] <jepler> cradek: unfortunately that's way before the firmware got put into CVS :(
[03:12:18] <skunkworks> 1/2 turn - 2122 and 8
[03:12:21] <Jymmmm> Hmmm, Is there some "encoder" thingy that maintains the current position even after a power off?
[03:12:39] <SWPadnos> now move it slowly - you get 2122, 8, ??, ?? .....
[03:12:40] <jepler> cradek: the firmware source, that is
[03:12:46] <skunkworks> 3/4 3169 and 12
[03:12:59] <Jymmmm> skunkworks 0 to 8000 ?
[03:13:05] <Jymmmm> min/max ?
[03:13:11] <skunkworks> 1 turn 4043 and 15
[03:13:22] <jmkasunich> number 2 = number 1 / 256
[03:13:22] <jepler> skunkworks: try loading the module with 'epp_wide=0'
[03:13:28] <cradek> well crap
[03:13:46] <jepler> 2122/16 = 8; 3169/256 = 12
[03:14:01] <jepler> skunkworks: (add epp_wide=0 to the loadrt line)
[03:14:17] <jmkasunich> jepler: huh? 2122/256 = 8
[03:14:44] <SWPadnos> 2122 mod 256 = 8
[03:15:04] <jepler> I meant 2122/256
[03:15:06] <SWPadnos> ah - typo
[03:15:09] <jepler> thinko
[03:15:13] <skunkworks> jepler: same thing
[03:15:28] <Jymmmm> E=mc2
[03:15:38] <Jymmmm> c3po
[03:15:43] <Jymmmm> r2d2
[03:16:02] <skunkworks> (edited the hal file and changed the line to epp_wide=0)
[03:16:08] <skunkworks> reran
[03:16:20] <jmkasunich> jepler: what exactly is READ() doing? emulating EPP mode?
[03:16:35] <jmkasunich> duh, yes
[03:16:47] <jmkasunich> !epp_soft
[03:16:59] <jmkasunich> didn't realize you had a software mode
[03:17:45] <jepler> it doesn't work
[03:17:53] <jmkasunich> what doesn't?
[03:17:54] <skunkworks> at some point it will just loose it and I can't get it back to aprox 0 by turning the opposite direction.
[03:17:55] <jepler> the epp soft mode
[03:18:20] <skunkworks> the numbers will jump by thousands ( not seeing the patern)
[03:18:30] <Jymmmm> skunkworks: Are you able to get the first three steps/increments by chance?
[03:18:50] <skunkworks> nope - just did it again - one turn is 4061 and 15
[03:18:51] <SWPadnos> skunkworks, the pattern people are thinking of is (number) and (number >> 8)
[03:19:05] <SWPadnos> (number >> 8 ) was that last (stupid smilies)
[03:19:22] <SWPadnos> right - 16 * 256 = 4096, 4061 is slightly belos that so you only get 15
[03:19:28] <SWPadnos> below
[03:19:33] <Jymmmm> skunkworks Sorry, I don't mena a FULL turn, just the the incrments, like 10,20,30 etx
[03:19:49] <jmkasunich> turn it as slowly as humanly possible ;-)
[03:19:51] <skunkworks> what I am saying is somewhere around 3 turns the number goes negative 4058
[03:19:56] <skunkworks> -4058
[03:20:09] <Jymmmm> what jmkasunich said, or get a LONG shaft and move just the end of it
[03:20:14] <Jymmmm> err handle
[03:20:18] <jmkasunich> I'm not surprised
[03:20:38] <jmkasunich> if it can't deal with a quarder turn correctly, there's no way its gonna deal with 3 turns correctly
[03:20:51] <skunkworks> :)
[03:20:58] <Jymmmm> how mayn steps per turn is it?
[03:21:02] <skunkworks> but it still turns back to 0
[03:21:05] <jmkasunich> sounds liek 4096
[03:21:09] <skunkworks> yes
[03:21:20] <skunkworks> 1024 I think line
[03:21:30] <Jymmmm> skunkworks can you move it three steps and get the value for each of the three steps?
[03:21:46] <skunkworks> 3 steps?
[03:21:53] <jmkasunich> jepler: if epp_soft doesn't work, then turning off epp_wide won't work either
[03:21:59] <SWPadnos> 3/4000 of a turn :)
[03:22:04] <jmkasunich> because that does the same thing
[03:22:16] <Jymmmm> skunkworks attach a LONG stick (handle) =)
[03:22:46] <jmkasunich> time out: are we beating on skunkworks because he's the only one who has the proble?
[03:22:54] <jmkasunich> or has jepler and/or cradek replicated it?
[03:23:17] <skunkworks> good question
[03:23:38] <jmkasunich> if they've replicated it, they can do far more troubleshooting without you doing anything
[03:23:54] <jepler> I haven't gotten my pluto out yet
[03:23:56] <skunkworks> I just like spinning the shaft
[03:23:59] <jepler> cradek had some kind of problem
[03:24:02] <jepler> but it sounded different
[03:24:04] <Jymmmm> * Jymmmm lol @ skunkworks
[03:24:12] <jepler> jmkasunich: are you sure? epp_soft and epp_wide should do different things
[03:24:26] <jmkasunich> if(epp_wide && !epp_soft)
[03:24:26] <skunkworks> the flashing between 2 differnt number was the odd thing in halmeter
[03:24:26] <jmkasunich> return inl(ioaddr+4);
[03:24:26] <jmkasunich> a = READ();
[03:24:28] <jmkasunich> b = READ();
[03:24:30] <jmkasunich> c = READ();
[03:24:32] <jmkasunich> d = READ();
[03:24:58] <jepler> int READ(void) {
[03:24:58] <jepler> if(epp_soft) {
[03:25:01] <jepler> } else
[03:25:01] <jepler> return inb(ioaddr+4);
[03:25:13] <jepler> so for epp_wide = 0, epp_soft = 0 it's supposed to use a series of inbs instead of one inl
[03:25:15] <jmkasunich> duh
[03:26:08] <jmkasunich> didn't look closely enough at READ, misread it as READ_SOFT
[03:26:12] <Jymmmm> jmkasunich That's your second "duh", one more and you have to down 3 shots of whiskey, 3 shots of tequilia, and 3 shots of vodka.
[03:26:21] <jmkasunich> no I don't
[03:26:29] <Jymmmm> Yep
[03:26:35] <jmkasunich> duh, duh, duh
[03:26:38] <jmkasunich> ;-P
[03:26:46] <Jymmmm> Line em up!
[03:27:34] <rayh> you missed one chug
[03:27:42] <Jymmmm> rayh that one is for you
[03:28:34] <rayh> Ah okay
[03:28:37] <rayh> done
[03:28:53] <skunkworks> should I get the latest build?
[03:29:17] <skunkworks> * skunkworks is feeling a bit useless here
[03:29:24] <Jymmmm> Hey, isn't mariss' usb thing called a rabbit?
[03:29:30] <SWPadnos> no
[03:29:30] <jmkasunich> G-rex
[03:29:31] <Jymmmm> http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=321630
[03:29:38] <SWPadnos> the microprocessor module he uses is called a rabbit
[03:29:38] <jmkasunich> it uses a rabbit processor board tho
[03:29:58] <SWPadnos> yep - that's the one
[03:30:15] <SWPadnos> that's not bad for the dev kit
[03:30:17] <Jymmmm> oh, just came across that at random.
[03:30:33] <Jymmmm> quadrature encoder inputs, PWM outputs, and pulse capture and measurement capabilities.
[03:30:45] <SWPadnos> piece of shit software
[03:30:51] <Jymmmm> ah
[03:30:52] <Jymmmm> General Purpose I/O: 33 parallel digital: 31 configurable, 2 fixed outputs
[03:31:12] <SWPadnos> proprietary, Windows only, not-quite-standard C, not-quite-standard Z80 CPU ...
[03:31:17] <Jymmmm> ew
[03:31:24] <SWPadnos> other than that, they're fine
[03:31:53] <Jymmmm> 0.22 optoisolator
[03:31:59] <Jymmmm> $0.22
[03:31:59] <SWPadnos> well, it's been fun, but I gotta get to bed.
[03:32:03] <Jymmmm> G'Night SWPadnos
[03:32:09] <SWPadnos> see you guys in a week or so, unles sI can find an internet cafe
[03:32:15] <jmkasunich> goodnight
[03:32:22] <Jymmmm> SWPadnos Florida?
[03:32:24] <rayh> See you Steven
[03:32:25] <SWPadnos> good-month ;)
[03:32:35] <SWPadnos> Jymmmm, yep - 6AM flight tomorrow :(
[03:32:47] <Jymmmm> have a good flight
[03:32:51] <SWPadnos> at least it'll be warm
[03:35:23] <skunkworks> the lite still flashesl
[03:36:29] <Ziegler> night everyone... thanks for the help
[03:37:27] <jepler> I'm simulating the verilog using "icarus" and I can't get the quadrature counter to misbehave
[03:37:55] <jmkasunich> are you also simulating the EPP transfers?
[03:38:03] <jmkasunich> cause thats what this smells like
[03:38:17] <jmkasunich> sometimes it winds up a byte off, and things get scaled by 256 because of it
[03:38:50] <jepler> no, I'm only testing the quadrature counter
[03:39:01] <jmkasunich> I bet its working fine
[03:39:16] <jmkasunich> because if you ignore the funky low numbers, the counts are correct
[03:39:38] <skunkworks> when I start going negative I get -32 and 63
[03:39:44] <jmkasunich> there's no way the counter itself is going 1000,4,1001,4,1003,4, etc
[03:39:54] <cradek> my behavior is different though - the count only shows 0, 1, or 2
[03:40:23] <jepler> cradek: does it change fast like it's a broken low 2 bits?
[03:41:07] <jepler> cradek: did you re-wire your setup for the new pinout? That's one thing that would have changed since the earliest versions
[03:41:25] <cradek> oh ...
[03:41:32] <cradek> probably not.
[03:41:33] <cradek> sorry.
[03:41:34] <jepler> goodnight guys -- hopefully I can set up my own pluto this weekend
[03:41:54] <cradek> goodnight
[03:41:57] <jepler> skunkworks: thanks for your report, sorry the software is the usual crummy jepler quality
[03:41:59] <jmkasunich> goodnight
[03:41:59] <cradek> did I give you that encoder?
[03:42:04] <jepler> I try to keep to "barely working", not "obviously broken"
[03:42:14] <cradek> you ought to have one to test with
[03:42:21] <jepler> cradek: no, but I have 2 servos with encoders, the lathe spindle encoder, and the idea of looping back digital outputs
[03:42:25] <jepler> I can come up with something
[03:42:30] <cradek> ok
[03:42:34] <jmkasunich> looping back is far better imo
[03:42:46] <jepler> I could also loop back in the verilog
[03:42:48] <jmkasunich> cause you can make it turn arbitrarily slow
[03:43:17] <cradek> goodnight jeff
[03:43:35] <skunkworks> jepler: no problem - it will work in time.
[03:45:31] <skunkworks> I think I need to send jepler some flowers - he seems down.
[03:46:10] <jepler> skunkworks: don't worry about me too much -- it's just my normal wry sense of humor
[03:46:11] <cradek> I'll buy him a drink tomorrow night, but I'm definitely not getting him flowers
[03:46:20] <skunkworks> :)
[03:46:21] <jmkasunich> lol
[03:47:13] <skunkworks> I was just happy I was getting 'something' back from the encoder :) I really figured it was something I was doing wrong.
[03:47:46] <cradek> it seems like we still don't know for sure, since mine is miswired so I can't test it
[03:48:07] <skunkworks> so there is hope still that it is me? ;)
[03:48:14] <jmkasunich> I bet its the data transfer from FPGA to PC
[03:48:15] <cradek> could be your hardware.
[03:48:37] <skunkworks> I could try a computer at work tomorrow
[03:49:08] <skunkworks> this is my portable.
[03:58:19] <rayh> I just found a copy of rt09j on a very old floppy disk
[03:58:54] <cradek> a very old what? :-)
[03:59:25] <rayh> Early 1999 so it was a later version of that patch.
[04:00:06] <rayh> How you doing tonight, Chris?
[04:00:29] <cradek> pretty decent I guess - ready for the weekend
[04:00:46] <cradek> I've been working hard these last couple weeks (I'm not used to it)
[04:01:14] <jmkasunich> ;-)
[04:01:25] <jmkasunich> working hard at work? or house stuff?
[04:01:27] <rayh> I've even got dos disks for servotogo
[04:01:37] <rayh> VBTune
[04:01:42] <cradek> at work - house stuff will always be behind
[04:04:17] <cradek> http://www.internationalpolka.com/images/RayHenry.jpg
[04:04:49] <jmkasunich> ;-)
[04:04:51] <skunkworks> the question is - how did you come across that?
[04:05:02] <jmkasunich> google I'd expect
[04:05:59] <rayh> been to a few polka fests but never played in one.
[04:06:41] <cradek> I prefer bluegrass but I can't play it either
[04:09:32] <skunkworks> Time for bed - thanks for all the help. Again - you guys are great.
[04:09:47] <cradek> goodnight
[04:51:48] <maddash> is there a guide/documentation on how to use hal_parport to **read** from the parallel port?
[04:54:36] <jmkasunich> do you mean using pins 2-9 for inputs?
[04:55:06] <jmkasunich> by default, pins 2-9 are outputs, plus four other outputs and five inputs
[04:55:16] <jmkasunich> total 12 out 5 in
[04:55:26] <jmkasunich> you can tell the driver to make 2-9 inputs
[04:55:32] <jmkasunich> total 4 out 13 in
[05:02:15] <maddash> hmmm what about 8-17?
[05:02:22] <jmkasunich> ?
[05:02:31] <jmkasunich> 8 is part of 2-9 ;-)
[05:02:56] <maddash> sry - take pin 17 for example
[05:03:53] <jmkasunich> goto page 52 of http://linuxcnc.org/docs/HAL_Documentation.pdf
[05:04:00] <jmkasunich> everything you ever wanted to know - with pictures even
[05:05:59] <maddash> mm, read that already, as well as the hal_* reference...but i'm not getting signals in halmeter when i ground pin 17
[05:06:58] <maddash> here's what i did: i wrote a .hal file for halcmd, and a program to create a pin and read from it every 5 sec
[05:07:01] <jmkasunich> do you have "addf parport.0.read <some-thread>" in your file?
[05:07:59] <jmkasunich> do a "show threads" (in halcmd) and look at the output
[05:08:10] <jmkasunich> if you don't have a parport read function in a thread, the port won't get read
[05:08:31] <maddash> http://pastebin.ca/318591 -- that's my src file
[05:08:38] <maddash> ooops, i meant http://pastebin.ca/368583
[05:08:58] <jmkasunich> that is C source
[05:09:16] <jmkasunich> has nothing to do with the question you asked
[05:09:21] <maddash> yes, of the program that is supposed to read from the pin
[05:09:53] <jmkasunich> but you asked about pin 17 of the parport
[05:10:16] <jmkasunich> is your parport working or isn't it?
[05:10:28] <jmkasunich> hint: you do NOT use your program to answer that question
[05:10:37] <jmkasunich> use halcmd show pin, or halmeter
[05:10:52] <jmkasunich> then, is your program working or isn't it -
[05:10:59] <jmkasunich> again, you do NOT use parport to answer that
[05:11:04] <jmkasunich> use setp
[05:11:27] <jmkasunich> you can test the parport and your program individually before you ever try to connect them together
[05:11:56] <maddash> i can **write** to pin 17 fine - this is confirmed with halmeter, and when i attach a voltometer between pin 17 and ground
[05:12:03] <jmkasunich> that is a key feature of hal - don't build complicated things until you know the simple building blocks work
[05:12:33] <jmkasunich> you didn't look at the pictures in the HAL manual
[05:12:42] <jmkasunich> pin 17 is ALWAYS an output
[05:13:08] <jmkasunich> if you want to test an input, connect to pin 13 or something
[05:14:09] <maddash> ok, i most likely read doc incorrectly
[05:14:39] <maddash> but with regards to creating a seperate thread in the .hal file - i never did this.
[05:15:21] <jmkasunich> loadrt threads name1=foo period1=1000000
[05:15:29] <jmkasunich> that creates a 1mS thread named foo
[05:15:32] <maddash> there was no seperate thread when i outputted to the parallel port; should this not be the same for input as well?
[05:15:47] <jmkasunich> wait
[05:16:04] <jmkasunich> are you doing this with a standalone HAL, or with EMC running?
[05:16:17] <maddash> standalone hal.
[05:16:32] <jmkasunich> then I have no idea how you got it to output
[05:16:42] <jmkasunich> did you load the parport driver with "loadrt" ?
[05:17:10] <maddash> yes. done inside the .hal file i fed to `halcmd -f [,,,]`
[05:17:33] <jmkasunich> did you have a "addf parport.0.write <something>" line in there?
[05:18:12] <jmkasunich> instead of me playing 20 questions, post the .hal file at pastebin
[05:18:13] <maddash> no. aside from the loadrt cmd, i also had a few linksp cmds
[05:19:41] <jmkasunich> btw, your program has a bug - very unlikely to happen, but...
[05:20:03] <jmkasunich> if the hal_malloc() fails, you need to call hal_exit() before you call exit(1)
[05:20:49] <jmkasunich> once you have successfully called hal_init(), you MUST call hal_exit before your program exits, or you will leave things allocated, etc
[05:21:21] <maddash> mmmm. actually, this program is just a hacked up version of one that jepler gave me - http://pastebin.ca/368583
[05:21:38] <maddash> oops. jepler's source is @ http://pastebin.ca/318591
[05:21:59] <jmkasunich> well, his is wrong too ;-)
[05:22:12] <jmkasunich> its a quick and dirty program, and thats ok
[05:22:34] <jmkasunich> hal_malloc rarely ever fails (you'd have to have about a thousand HAL pins already allocated, for example)
[05:22:44] <jmkasunich> but a real program needs to properly handle errors like that
[05:22:54] <jmkasunich> btw, how do you exit from that program? ctrl-C?
[05:23:11] <jmkasunich> if so, you don't call hal_exit - same problem
[05:23:21] <maddash> no, i call `halcmd unload all`
[05:23:30] <jmkasunich> ok, that calls hal_exit for you
[05:23:42] <jmkasunich> I think
[05:23:53] <maddash> if i press ctrl+c, i cannot reload the program, until i kill off realtime
[05:24:06] <maddash> ie, `scripts/realtime stop`
[05:24:09] <jmkasunich> yeah, thats the leftover stuff I was telling you about
[05:24:28] <jmkasunich> if you did a halcmd show after doing ctrl-c, the program's pin might still be there
[05:25:11] <jmkasunich> a robust program would have a signal handler to intercept ctrl-C and call hal_exit() before quitting
[05:25:22] <maddash> darn. i just realized that my ssh access to the machine i was working on (the one with emc's hal on it) is blocked off...i don't have the .hal file here, but i could try to duplicate it form memory
[05:25:33] <jmkasunich> thats ok
[05:26:01] <jmkasunich> I'm in the middle of something - if you aren't in a position to actually try stuff now, it doesn't make much sense to talk about it
[05:27:50] <maddash> ok, i'll get back to you later
[05:28:02] <maddash> thanks, anyway
[07:51:13] <sebastienbailard> /part
[08:03:35] <CIA-6> 03jmkasunich 07TRUNK * 10emc2/src/hal/utils/halcmd_main.c: fix readline support in halcmd
[08:08:11] <K`zan> Night all
[08:08:23] <jmkasunich> goodnight
[09:31:21] <jmkasunich_> jmkasunich_ is now known as jmkasunich
[10:53:54] <fabiotv> I need to move 2 axis, 1 linear "X" and 1 angular "A" according to 2 switches "B" and "C" normal open. with parallel port 378. If switch "B" is closed and switch "C" is open move only axis "X", if switch "B" is closed and switch "C" closed move axis "X" and axis"A" 90° . As I can make ?!?!
[11:45:00] <anonimasu> http://www.tuxcnc.org/pivot/entry.php?id=14#body
[11:45:00] <anonimasu> yay!
[11:47:42] <anonimasu> I wish he would tell us about the bugs
[12:26:37] <fabiotv> for bugs what you mean?
[12:30:03] <anonimasu> the bugs paul mentions on the page?
[13:03:46] <fabiotv> I have not found nothing in these pages!
[13:04:29] <anonimasu> what?
[13:07:27] <fabiotv> I need to move 2 axis, 1 linear "X" and 1 angular "A" according to 2 switches "B" and "C" normal open. with parallel port 378. If switch "B" is closed and switch "C" is open move only axis "X", if switch "B" is closed and switch "C" closed move axis "X" and axis"A" 90° . As I can make ?!?!
[13:08:30] <fabiotv> as I make to read the switch to the gcode
[13:09:08] <anonimasu> I have no idea what you mean
[13:09:10] <anonimasu> sorry :(
[13:11:17] <fabiotv> is impossible or you have not understood what I mean?
[13:12:08] <anonimasu> it's possible, but I dont understnd what you mean
[13:15:32] <fabiotv> What is not clear for you?
[13:17:56] <anonimasu> what you mean
[13:19:08] <skunkkworksz> skunkkworksz is now known as skunkworks
[13:23:06] <fabiotv> I try to explain better i have to move 2 axis one linear called X and one angular called A; this axis begin to move only when the switches are opened or closed.
[13:25:33] <anonimasu> I think you should wait around until somone better then me comes by here..
[13:29:42] <skunkworks> sounds like paul almost has something to release.. good for him.
[13:31:30] <anonimasu> skunkworks: yeah, now if he'd just fsck off, and stop being so concerned with emc2..
[13:33:19] <anonimasu> :)
[13:33:32] <skunkworks> everyone need a hobby
[13:34:22] <anonimasu> yeah spreading FUD, is a nice one..
[13:36:13] <jepler> fabiotv: you cannot read the state of external switches from g-code.
[13:36:33] <xemet> hello
[13:36:40] <skunkworks> Hi
[13:36:55] <xemet> anyone knows how to do that?
[13:36:56] <xemet> http://axis.unpy.net/files/01171767993/spline-velocity-profile.png
[13:37:03] <xemet> is that Halscope?
[13:37:08] <skunkworks> I bet jepler does ;)
[13:37:33] <xemet> me too
[13:37:38] <skunkworks> jepler: I plan to set up a computer here at work to try out the pluto (different hardware)
[13:38:39] <jepler> skunkworks: OK -- that will tell us something, but I'm not sure what
[13:39:01] <jepler> xemet: yes, that's halscope
[13:39:27] <xemet> uhm...and how can I read the X or Y vel?
[13:39:37] <xemet> or XY...
[13:39:44] <skunkworks> jepler: you don't think it could just be my portable being stupid?
[13:40:00] <xemet> I have the pluto
[13:40:19] <xemet> my pluto had a problem with the parallel port of my laptop
[13:40:29] <xemet> is that the case?
[13:40:36] <jepler> xemet: take a look at core_sim.hal -- it uses 'ddt' to find the speed of each axis independently, then 'hypot' to find the speed of XY together, and then the speed of XYZ
[13:40:37] <skunkworks> worked fine on a workstation?
[13:41:16] <skunkworks> xemet: what did it do?
[13:42:28] <xemet> it didn't config
[13:43:06] <skunkworks> xemet: are you reading encoders back into it?
[13:43:29] <xemet> no at the moment...I've got it the last week
[13:43:39] <xemet> and still not having time to try somthing
[13:43:50] <skunkworks> I have the same problem ;)
[13:43:51] <xemet> I tested only if it works using the sample configs
[13:44:18] <xemet> at the beginning it didn't config and the problem was the parallel port of my laptop
[13:44:30] <xemet> I had to change it on bios
[13:44:39] <xemet> from bi-directional to only output
[13:45:05] <skunkworks> hmm - I have a setting in the bios for epp. That is what it is set to right now.
[13:46:25] <xemet> and you are configuring it with the EPP option?
[13:46:34] <xemet> have you installed a driver?
[13:46:53] <skunkworks> I am using the hal pluto componant.
[13:47:20] <xemet> ah...no, I've tested it on windows...
[13:47:41] <jepler> I never got any of the epp-mode stuff to work on windows with my laptop
[13:47:45] <skunkworks> ah - ok. have not done that here - strictly just use the hal stuff
[13:47:47] <jepler> only the basic SPP-mode stuff worked
[13:52:02] <fabiotv> xemet: You have one email in hotmail
[13:52:39] <xemet> fabio, un attimo che sono al telefono
[13:53:43] <fabiotv> ok aspetto!!
[13:58:09] <duerz> hello, i am interested in trying this cnc software out - but no nothing about linux. i have some questions
[13:59:05] <jepler> duerz: ask away
[13:59:07] <duerz> Will this emc run on solaris 10?
[14:00:01] <jepler> no, it is for x86 linux with "real-time" kernel patches only. the precompiled packages are for ubuntu 6.06.
[14:01:54] <jepler> bbl
[14:07:25] <xemet> eccomi
[14:07:49] <fabiotv> bene
[14:08:06] <xemet> forse è meglio se ci trasferiamo...hai msn?
[14:08:09] <duerz> is this x86 linux available for free download?
[14:08:21] <xemet> duerz: of course
[14:08:41] <xemet> www.ubuntu.com
[14:08:42] <fabiotv> no, sono con ubuntu
[14:08:52] <xemet> hai gaim?
[14:09:03] <fabiotv> guardo?
[14:09:12] <xemet> dovrebbe esserci, nel menu internet
[14:10:03] <fabiotv> trovato, devo fare account e password
[14:10:39] <xemet> uhm...non hai nessun account msn aim..icq o altro?
[14:10:42] <duerz> ok what is the name of file i download from this site.?
[14:11:22] <xemet> duerz: do you need ubuntu with EMC2 or only Ubuntu?
[14:11:39] <fabiotv> no!!
[14:12:01] <duerz> i need the linux software and emc2 - i think i can get the emc 2 from the web site though
[14:12:33] <skunkworks> jepler: this 450 here has epp and 2 versions - 1.7 and 1.9. does it matter?
[14:12:38] <skunkworks> 450mhz
[14:12:52] <xemet> duerz: no, download the the live CD that has EMC2 pre-installed
[14:12:55] <xemet> http://linuxcnc.org/iso/emc2-ubuntu6.06-desktop-i386.iso
[14:13:11] <xemet> fabio: vabè parliamo qui
[14:13:27] <xemet> non so se si può paralre in italiano o ci cacciano...
[14:13:32] <xemet> :)
[14:14:04] <duerz> thank you xemet :)
[14:14:13] <xemet> welcome
[14:14:14] <fabiotv> proviamo!
[14:14:27] <xemet> ok...allora ho letto la tua email. Ho solo un dubbio
[14:14:32] <duerz> next question - can i use isagraph for the logic - st, sfc?
[14:14:45] <xemet> duerz: I don't know...sorry
[14:15:05] <fabiotv> speak me
[14:15:19] <xemet> fabio: l'asse X e l'altro si devono muovere quando premono gli switch durante la lavorazione?
[14:15:40] <xemet> perchè dici che devi leggere lo switch con un g-code?
[14:16:36] <fabiotv> da quando leggono gli switch partono a fare la lavorazione
[14:16:59] <xemet> allora...quindi è un processo da fare prima, diciamo come l'homing?
[14:17:13] <fabiotv> qualcosa del genere
[14:17:19] <xemet> ok, si può fare
[14:17:27] <xemet> non credo sia neanche complicato...forse
[14:17:27] <fabiotv> come
[14:17:35] <xemet> adesso ne so un po' più di hal...
[14:17:55] <xemet> allora...ci devo studiare su, ma fondamentalmente mi pare che
[14:18:21] <xemet> si dovrebbe aggiungere un pulsante all'interfaccia (come se avessi un latro pulsante homing) per far partire la procedura
[14:18:33] <xemet> e questo si può fare con pyVCP
[14:18:55] <xemet> dopo di che creare un piccolo modulo in HAL che legga gli switch e comandi i movimenti a seconda di ciò che arriva
[14:19:10] <xemet> in linea di massima dovrebbe essere qualcosa del genere
[14:20:30] <fabiotv> praticamente se mi legge uno switch mi deve fare una rotazione di 90°
[14:20:57] <xemet> allora...
[14:21:02] <xemet> hai motori stepper?
[14:21:10] <fabiotv> si
[14:21:23] <fabiotv> mezzo passo
[14:21:57] <xemet> quando tu avvii emc2 si avvia un modulo HAL stepgen
[14:22:04] <xemet> che genera gli step del motore
[14:22:10] <xemet> per comandare la posizione
[14:22:23] <fabiotv> si
[14:22:36] <fabiotv> quello e tutto apposto
[14:22:46] <xemet> il modulo stepgen viene comandato normalmente dal G-code quando avvii la lavorazione
[14:22:59] <xemet> noi lo comandiamo prima
[14:22:59] <fabiotv> si
[14:23:11] <fabiotv> come
[14:23:16] <xemet> e gli facciamo generare i movimenti necessari a fare la rotazione o quant'altro
[14:23:24] <xemet> a seconda di ciò che fanno i pin degli interruttori
[14:23:34] <fabiotv> ok
[14:23:37] <xemet> proprio ieri ho imparato a fare piccoli moduli in hal
[14:24:01] <xemet> anzi...lasciando stepgen...dovrebbe essere possibile usare i counts per il jog...direttamente per muovere i gli assi
[14:24:28] <xemet> aspetta un secondo che apro EMC2 sull'altro computer
[14:25:14] <duerz> does anyone know if one can use anything other than the ladder within emc2? isagraph?
[14:25:16] <fabiotv> prima una cosa, si può configurare un asse lineare e uno angolare?
[14:25:29] <xemet> si
[14:25:59] <xemet> non l'ho mai fatto perchè ho solo 3 assi...
[14:26:08] <fabiotv> perchè ho provato ma qualcosa non gira
[14:26:11] <xemet> ma è sicuramente possibile
[14:26:32] <xemet> nel file ini non puoi sceglier il tipo di asse tra linear e angular?
[14:26:42] <fabiotv> si ok
[14:27:04] <xemet> cos'è che non gira?
[14:27:55] <fabiotv> ho provato a togliere due assi alla configurazione xyza ottenendo solo x e a
[14:28:12] <xemet> hai il file a portata di mano?
[14:28:25] <fabiotv> si
[14:28:40] <fabiotv> non legge i gradi
[14:28:41] <xemet> copia il contenuto e incollalo qui:
[14:28:42] <xemet> http://www.pastebin.ca/
[14:28:45] <xemet> e poi mandami il link
[14:29:21] <fabiotv> solo file .ini
[14:29:51] <xemet> prima l'ini
[14:30:00] <xemet> poi vediamo se serve altro
[14:32:52] <xemet> ci sei??
[14:33:22] <fabiotv> si
[14:33:34] <xemet> l'hai messo su pastebin?
[14:34:14] <fabiotv> quello con l'asse angolare l'ho buttato, ho solo questo che ci sto lavorando
[14:34:38] <xemet> comunque, fondamentalmente non devi cambiare solo il file ini
[14:34:44] <xemet> ma anche il file hal
[14:35:05] <fabiotv> il link dove
[14:35:30] <xemet> quale link?
[14:35:50] <fabiotv> il link pastebin
[14:35:57] <xemet> http://www.pastebin.ca/
[14:36:03] <xemet> quando incolli qualcosa e poi clicchi su submit
[14:36:08] <xemet> ti dovrebbe dare un link
[14:36:09] <duerz> xemet - do you know of anyone who might be able to answer my question?
[14:36:47] <fabiotv> http://www.pastebin.ca/369004
[14:36:49] <xemet> duerz: a lot of people...but maybe they are not reading at the moment
[14:37:42] <xemet> duerz: you could ask in the mailing list...
[14:37:49] <xemet> fabio un attimo che me lo leggo
[14:38:42] <skunkworks> duerz: as far as I know you have classicladder and hal componats. (you could write your own hal componants also)
[14:38:49] <skunkworks> componants
[14:39:23] <fabiotv> niente di particolare, gira
[14:40:33] <xemet> per curiosità, perchè hai usato il file demo_stepper invece di quello stepperxyza?
[14:41:12] <fabiotv> perchè c'è classicladder
[14:41:20] <xemet> ah ok
[14:41:42] <xemet> ah si...infatti si chiama CL...non avevo visto
[14:41:58] <xemet> ma questo funziona?
[14:42:06] <fabiotv> si
[14:42:16] <xemet> e quindi dov'è il problema?
[14:42:32] <xemet> ma hai modificato anche il file demo_step_cl.hal?
[14:42:48] <fabiotv> se gli dico di fare 90° o di fare 20mm non cambia
[14:42:53] <fabiotv> si
[14:43:36] <xemet> eh...
[14:43:38] <xemet> certo
[14:43:42] <xemet> devi metter angular
[14:43:55] <xemet> gli assi sono tutti e due linear
[14:45:23] <fabiotv> si
[14:45:32] <xemet> hai provato?
[14:45:41] <fabiotv> si
[14:45:45] <xemet> e non va?
[14:45:51] <fabiotv> si
[14:45:57] <xemet> va o non va??
[14:46:08] <fabiotv> non va
[14:46:21] <xemet> uhm...
[14:46:33] <xemet> fammi capire
[14:46:58] <xemet> il tutto è attaccato alla macchina o stai solo simulando?
[14:46:58] <fabiotv> prova a modificare stepperxyza togliendo yz e poi vedi
[14:47:17] <fabiotv> collegato i motori
[14:47:36] <xemet> allora aspetta un attimo...mi trasferisco sull'altro computer e provo
[14:47:59] <fabiotv> posso fumare una cicca?
[14:48:12] <xemet> :) non penso sia vietato su emc
[14:48:54] <fabiotv> fumo fuori, solo 5 minuti
[14:49:15] <duerz> is this italian :)
[14:49:37] <xemet> duerz: yes
[14:49:59] <duerz> i used to work for fidia - not long ago
[14:51:13] <duerz> spent some time in turin, also - loved samara's :)
[14:55:15] <fabiotv> se mi mandi via email il tuo numero di telefono ti chiamo!
[14:55:37] <fabiotv> cosi ci capiamo meglio
[15:09:12] <xemet> allora...almeno ho riprodotto il problema
[15:09:23] <xemet> Non mi legge l'asse A...ma sempre Y
[15:11:37] <fabiotv> esatto
[15:12:24] <fabiotv> non guardare i parametri di accelerazione ecc. è tutto da mettere apposto
[15:13:25] <xemet> aspetta che chiediamo ai guru
[15:13:30] <xemet> jepler are you therE?
[15:15:57] <fabiotv> in [TRAJ] ai messo DEFAULT_ANGULAR_VELOCITY and MAX_ANGULAR_VELOCITY
[15:17:04] <xemet> ho lasciato quelle che c'erano
[15:17:31] <xemet> ma non è quello il problema...
[15:17:38] <xemet> io ho chiamato gli assi X e A
[15:17:44] <xemet> quando avvio EMC
[15:17:55] <xemet> se scrivo G0 A10 l'asse A non si muove
[15:18:13] <fabiotv> anche a me!
[15:18:16] <xemet> mentre se scrivo G0 Y10 (Y che non dovrebbe esistere) si sposta l'asse A
[15:18:27] <fabiotv> esatto
[15:19:14] <fabiotv> è veramente strano
[15:19:16] <xemet> secondo me c'è un problema nel file hal...ma non capisco ancora cosa
[15:20:35] <fabiotv> forse ci sono delle impostazioni di base, per la grafica ?!?
[15:21:30] <xemet> no
[15:21:53] <xemet> la grafica deve rimanere uguale
[15:22:20] <fabiotv> non so!
[15:23:30] <xemet> jepler...are you there?
[15:23:40] <xemet> jepler sa tutto...selo troviamo si risolve
[15:24:10] <xemet> allora faccio un altra prova
[15:24:41] <fabiotv> se fossero tutti e due angolari cosa fà!
[15:26:44] <xemet> è possibile mettere gli assi come si vuole
[15:26:53] <xemet> su questo non c'è dubbio
[15:27:17] <xemet> ma non va il link tra l'asse A che ho e il programma che legge il gcode
[15:27:30] <xemet> infatti A spunta come coordinata
[15:27:44] <xemet> ma in realtà il secondo asse non è A ma sempre Y
[15:28:09] <xemet> infatti se scrivo G0 Y10 si muove lungo Y e mi cambia il valore di A
[15:28:39] <xemet> cmq il problema si risolve sicuramente
[15:28:52] <xemet> c'è sicuramente qualcosa che non sappiamo
[15:29:06] <fabiotv> e si comporta da asse lineare e non angolare
[15:30:39] <fabiotv> e questo lo fà anche nella versione DISPLAY AXIS
[15:30:49] <xemet> si io uso quella
[15:32:17] <fabiotv> e anche nella versione DISPLAY TKEMC
[15:32:46] <xemet> non riesco bene a capire nel file hal come fa lui a sapere che axis.0 o axis.1 devono corrispondere agli assi 1 e 2 del file ini
[15:33:03] <xemet> ho provato addirittura a cambiare tutti gli assi axis.1 con axis.3
[15:33:12] <xemet> e continua a funzionare in quel modo
[15:33:29] <xemet> ma nel file ini ho solo axis_0 e axis_1
[15:33:47] <xemet> quindi con axis.3 mi dovrebbe dare un errore...
[15:34:34] <fabiotv> in DISPLAY AXIS sulla grafica scrive asse A, ma dove c'e il punto da selezionare con il mouse è sempre Y
[15:36:09] <fabiotv> con il turbocnc sono riuscito a mettere tutto apposto assi angolari switch ecc. ma emc2 niente
[15:38:10] <xemet> ti dico che al 100% funzionerà
[15:38:17] <xemet> entro quando ti serve?
[15:39:06] <xemet> io credo di intuire dove sia il problema
[15:39:18] <xemet> ma non capisco perchè non mi funziona
[15:39:44] <xemet> il problema è che lui numera gli assi in modo sequenziale
[15:40:02] <xemet> nel senso, se io gli do 1 solo asse, lui in automatico me lo assegna alla X
[15:40:40] <xemet> così se gliene do 2 lui me li ha assegnati alla X e alla Y
[15:41:20] <xemet> a questo punto mi viene effettivamente il dubbio se si possano collegare un solo asse lineare e uno rotativo...
[15:41:27] <xemet> ma mi sembra molto strano che non si possa fare
[15:41:46] <xemet> se solo ci fosse jeff epler già avremmo una risposta
[15:42:47] <fabiotv> facciamo che siano due assi lineari con gli switch come posso fare?
[15:43:17] <xemet> aspè squilla il tel
[15:43:30] <fabiotv> ok
[15:49:45] <xemet> allora
[15:50:02] <xemet> se fossero due lineari si farebbe come ho detto prima
[15:50:18] <xemet> ci si fa un componente hal che gestisce la situazione
[15:51:09] <fabiotv> del tipo
[15:51:20] <xemet> hai letto questa paginetta (sempre riguardo al joystick) che ho scritto ieri?
[15:51:22] <xemet> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?HowToWriteNewHalComponent
[15:51:50] <xemet> qui ho creato un componente hal semplicissimo che genera un contatore a partire da un inmput tipo float
[15:52:11] <xemet> ovviamente si possono fare cose molto più complesse che gestiscono tanti pin e tante condizioni
[15:53:11] <xemet> quindi, ci serve qualcosa che, per esempio, comincia a fare muovere un asse, quando becca l'interruttore si ferma, e comincia a fare muovere l'altro di un certo valore
[15:53:37] <xemet> insomma, puoi programmare più o meno qualsiasi cosa...
[15:55:10] <fabiotv> il classicladder non serve
[15:55:49] <xemet> probabilmente si potrebbe fare anche con quello...ma io non ho esperienza attualmente
[15:56:03] <xemet> se tui conosci i PLC e sai programmarli probabilmente potresti provare
[15:57:06] <fabiotv> ho solo provato classicladder per questa applicazione con scarsi risultati
[15:57:40] <fabiotv> se si può fare tutto con solo emc2 è meglio
[15:57:49] <xemet> credo di si.
[15:57:57] <xemet> entro quando ti serve questa cosa?
[15:58:20] <xemet> sai programmare in C?
[16:01:54] <xemet> ci sei?
[16:03:39] <fabiotv> non so programmare in c
[16:03:55] <fabiotv> quanto prima
[16:04:13] <xemet> uhm...neanche io, ma sto imparando
[16:04:17] <xemet> ho un bel libro
[16:04:26] <fabiotv> titolo
[16:04:34] <xemet> ma ancora poco tempo per leggerlo
[16:04:39] <xemet> è in inglese però
[16:04:47] <xemet> Software engineering in C
[16:06:16] <xemet> i componenti Hal si scrivono in C
[16:06:35] <xemet> ma non credo sia un problema perchè con la programmazione ho abbastanza dimestichezza
[16:06:52] <xemet> alla fine se si conoscono altri linguaggi uno nuovo non è troppo difficile da imparare
[16:07:32] <fabiotv> bisogna avere delle basi
[16:08:01] <fabiotv> di azionamenti mi capisco
[16:08:25] <fabiotv> ma mi perdo con 2 switch
[16:08:57] <xemet> io so programmare in python, pascal, sapevo il basic ma l'ho abbandonato perchè secondo me inutile, e programmo anche in actionscript (linguaggio interno a flash, tipo java)
[16:09:17] <xemet> il C mi manca e DEVO impararlo
[16:09:45] <xemet> cmq la logica di HAL è un po' diversa dalla normale programmazione, conosci il LabView
[16:09:56] <xemet> HAL è qualcosa di molto simile...solo senza grafica
[16:10:21] <fabiotv> solo sentito parlare
[16:10:30] <khiraly123> hi!
[16:10:53] <xemet> http://www.k-team.com/images/products/labview-wiring.gif
[16:10:56] <xemet> hi
[16:11:14] <xemet> qui c'è la grafica e vedi i componenti e i collegamenti...in hal no...
[16:12:20] <xemet> comunque...io mi sa che devo andare a studiare...se no non mi laureo più
[16:13:04] <xemet> il problema dei due assi lo risolviamo...stasera vedo se trovo Jeff epler e gli chiedo
[16:13:17] <xemet> e dopo risolviamo anche il problema degli switch
[16:13:32] <xemet> la tua email ce l'ho, non appena so qualcosa ti scrivo
[16:13:32] <fabiotv> ok grazie
[16:13:33] <xemet> ok?
[16:13:44] <fabiotv> buon studio
[16:13:48] <xemet> grazie
[16:13:53] <xemet> byeeeeee
[16:14:02] <khiraly123> can somebody help me in EMC?
[16:14:43] <khiraly123> I have downloaded the live cd .iso, booted from it, and started emc2 with the axis gui
[16:15:09] <fabiotv> ok good
[16:15:13] <khiraly123> (first issue: I have no pc speaker, and no beep sounds. On my desktop system (Debian sid) I have beeps)
[16:15:42] <khiraly123> SO I have choosed the stepper->stepper_mm option, and tried to run the axis.ngc file
[16:16:06] <khiraly123> It looks ok on screen, but there is no signal on pc parallel port (checked with an oscilloscope)
[16:16:27] <khiraly123> (it look ok on screen = there is an animation)
[16:17:03] <khiraly123> What did I miss?
[16:19:53] <khiraly123> for example on pin 3 there is a constant 5V output, but should have a square signal
[16:29:52] <fabiotv> par port 378 is ok?
[16:41:16] <khiraly123> parport 378? what do you mean? The 3,7,8 pin?
[16:42:46] <fabiotv> no parallel port lpt
[16:43:05] <tomp> khiraly123: 1st: which configuration did you choose when you 'started emc' ( 2nd: parport adddress 0x378 was referred to )
[16:44:29] <tomp> fabio: buon giorno, solo picolo Italiano, grazie
[16:44:54] <fabiotv> bene
[16:46:08] <tomp> khiraly123: when you run emc, you must choose a configuration from a dialog with a little penguin... which did you choose? OR you may be using an older emc.. please describe
[16:52:03] <tomp> khiraly123: please look at this http://imagebin.org/7365 you must have chose one from something like this. The root will vary a bit on your system, but the tree should look very similar. whcih did you choose?
[16:52:13] <jepler> khiraly123: one common reason that no signals appear on the parallel port is described here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Parallel_port_no_longer_works_in_EMC_2_0_1_or_later_emc_starts_but_motors_don_t_turn
[16:54:30] <tomp> fabiotv: thank you, nice to meet you, jepler, thanks, khiraly123: keep trying me: gotta go!
[16:56:46] <khiraly123> thx for your help
[16:56:56] <khiraly123> I have choosed the stepper->stepper_mm option
[16:57:19] <khiraly123> I use the ubuntu live .iso file (which is downloadable from the site)
[16:57:26] <khiraly123> jepler: Im reading, thx
[17:00:15] <khiraly123> It seems goo in the dmesg
[17:00:21] <khiraly123> I will paste on the net somewhere
[17:00:34] <jepler> http://pastebin.ca/
[17:08:43] <khiraly123> just a sec, I need a usb pen, to copy over the dmesg message (dont have net on the ubuntu live cd)
[17:15:33] <khiraly123> done
[17:16:44] <alex_joni> khiraly123: you also need to provide us the link on pastebin
[17:17:04] <khiraly123> yes, I know, just the webpage loading
[17:17:10] <khiraly123> sorry for the delay;-\
[17:17:19] <khiraly123> http://pastebin.com/887371
[17:18:18] <khiraly123> I dont see any problem in this log. I make some picture, and document the exact steps for you
[17:18:41] <khiraly123> a simple question: If the parport is not connected to anywhere, can I see the signals on it?
[17:18:43] <alex_joni> khiraly123: the default configuration (stepper_mm) assumes the parport is defined at address 0x378
[17:19:05] <alex_joni> you should make sure (in BIOS) that is the correct address
[17:19:45] <alex_joni> khiraly123: also, I see your machine has problems with RT (that's probably because of an onboard graphics adapter)
[17:19:54] <khiraly123> alex_joni: the bios is the only way to check? (hp nx9105)
[17:20:18] <khiraly123> aham. I have an nvidia geforce 420 go 32M onboard
[17:20:27] <khiraly123> (it is a laptop)
[17:20:39] <khiraly123> I check the bios. wait a sec
[17:20:54] <alex_joni> khiraly123: maybe lspci or cat /proc/ioports might help too
[17:21:09] <alex_joni> khiraly123: laptops are usually problematic on RT
[17:21:20] <khiraly123> rt problem is not because the screensaver on the ubuntu live?
[17:22:31] <alex_joni> khiraly123: it can be a dozen things
[17:22:37] <jepler> rt must work reliably 100% of the time, even while a screensaver is on
[17:22:56] <alex_joni> cradek in here, reported that RT had problems every time his CPU cooler went on/off
[17:23:36] <khiraly123> hmm. there is no info about the parport in the BIOS. (the only info is that: Parallel Port Mode: ECP/EPP
[17:24:23] <khiraly123> I'll try the /proc/ioports and lspci
[17:25:43] <khiraly123> I'll check if the powernowd module is loaded or not
[17:31:50] <khiraly123> no in in lspci
[17:32:59] <khiraly123> 0378-037b: hal_parport
[17:33:08] <khiraly123> ^ in the /proc/ioports
[17:33:42] <khiraly123> I assume it is good
[17:33:57] <khiraly123> thoughts?
[17:34:03] <alex_joni> khiraly123: that only says where the current driver allocated it
[17:34:17] <jepler> did you try the suggestion from the wiki and add probe_parport to your hal file?
[17:34:26] <alex_joni> jepler: does the pc_parport driver autodetect the parport?
[17:37:40] <khiraly123> jepler: is it possible in the live CD?
[17:38:21] <alex_joni> khiraly123: yes, if you copy the configs to your home dir
[17:38:25] <khiraly123> dpkg -l |grep emc2 says the version is 2.1.0
[17:38:34] <alex_joni> but you'll lose the changes next time you start the LiveCD
[17:38:56] <khiraly123> is it a midnight commander (mc) on the live cd somewhere
[17:39:14] <khiraly123> its a major pita navigating without mc :-(
[17:39:14] <alex_joni> khiraly123: unfortunately not
[17:39:22] <khiraly123> ok
[17:39:46] <alex_joni> khiraly123: you need to follow this: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CustomizingConfigsOnUbuntu
[17:40:53] <khiraly123> Im reading, thx
[17:42:47] <khiraly123> I have the emc2 dir in my ~ dir (the first question what ask the graphical gui). So I can customize
[17:43:16] <khiraly123> ~/emc2/configs/stepper/
[17:46:37] <khiraly123> ok, I have written the probe_parport, here is the dmesg:
[17:47:05] <khiraly123> parport: PnPBIOS parport detected, io_lo=378 io_hi=778
[17:47:18] <khiraly123> config string '0x0378'
[17:47:27] <khiraly123> PARPORT: installed driver for 1 ports
[17:47:29] <khiraly123> ennyi
[17:47:34] <khiraly123> thats it
[17:52:49] <khiraly123> On which pin should I see the square signals? on pin 2 for example?
[17:53:39] <jepler> in standard_pinout.hal, the X step signal is on pin 2
[17:54:10] <khiraly123> And If I connect nowhere, can I measure with an oscilloscope?
[17:54:56] <jepler> yes. while stepping you would expect to see the signal be HIGH for around 50 microseconds and LOW for a varying length of time.
[17:58:29] <jepler> ooh lunchtime!
[17:59:01] <Martin_Lundstrom> alex_joni, can your robots do this? http://video.google.com/videoplay?docid=2079889503361061502&q=type%3Agpick&hl=en
[18:01:59] <alex_joni> Martin_Lundstrom: :P
[18:02:25] <alex_joni> those people are crazy imo
[18:02:30] <alex_joni> I trust suftware .. but not that much :D
[18:05:45] <Martin_Lundstrom> quite powerful
[18:10:28] <xemet> hi
[18:10:58] <xemet> anyone there?
[18:11:36] <skunkworks> xemet: Thanks for helping out this morning..
[18:12:10] <skunkworks> have no clue what you two where talking about ;)
[18:12:25] <alex_joni> hi xemet
[18:13:11] <xemet> question from a friend of mine
[18:13:47] <xemet> skunk: now you will understand
[18:14:09] <xemet> is possible in Axis to set only two axes?
[18:14:16] <xemet> one linear and one angular?
[18:16:56] <khiraly123> http://i12.tinypic.com/44a1h1t.jpg
[18:17:17] <khiraly123> http://i9.tinypic.com/4gpcs2a.jpg
[18:17:42] <khiraly123> these two pictures shows how I wanted to check the parallel port pins
[18:21:53] <khiraly123> I will try with some different computers in the next few days.
[18:23:06] <xemet> someone got my question?
[18:24:28] <alex_joni> xemet: I saw it
[18:24:41] <alex_joni> it's possible for emc to control only 2 axes, one linear and one angular
[18:24:55] <alex_joni> but you probably want to write some kinematics to make it understand how that moves
[18:25:09] <alex_joni> I'm not sure what you want to display in AXIS though
[18:25:29] <xemet> uhm...I've tried the file stepper xyza
[18:25:59] <xemet> I would like to command only the X axis and the A axis
[18:26:09] <xemet> and remove Y and Z
[18:27:48] <xemet> Axis could continue to show XYZ...
[18:28:15] <xemet> but if I type G0 Y10 it hasn't to move
[18:28:30] <xemet> and If I type A10 it has to rotate 10 degrees
[18:28:58] <xemet> is it possible...I think it should if it works with four axis...why not work for just two?
[18:35:09] <alex_joni> xemet: because the trivial kinematics assignment works like this:
[18:35:24] <alex_joni> motor0 = X, motor1=Y, motor2=Z, motor3=A
[18:35:37] <xemet> ok, so I should write new kinematics...
[18:35:41] <alex_joni> so if you want X and A (and don't like to recompile) then define all 4
[18:35:45] <alex_joni> and only use X & A
[18:35:50] <xemet> big question
[18:35:54] <xemet> how to do that?
[18:35:59] <xemet> is it described in the manual?
[18:36:07] <alex_joni> xemet: how do you mean?
[18:36:14] <alex_joni> just use the xyza config
[18:36:19] <xemet> no
[18:36:36] <xemet> XA, writing new kinematiks and recompiling
[18:36:37] <alex_joni> you mean the kinematics?
[18:36:39] <xemet> yes
[18:36:47] <xemet> :)
[18:36:54] <alex_joni> I wrote a bit about kinematics here: http://www.linuxcnc.org/docs/2.1/html/motion/kinematics/index.html
[18:36:57] <xemet> if I've to learn, better to learn new things
[18:37:10] <xemet> I know how to use xyza...
[18:37:43] <xemet> Great
[18:37:49] <xemet> I will read it
[18:38:48] <alex_joni> xemet: but if you're g-code will have Y or Z in it, then you have a problem
[18:39:04] <xemet> of course...I understand
[18:39:17] <alex_joni> from my experience it will just hang waiting for the motion controller to reach a certain Y position, which it never will (as there's no Y movement)
[18:39:18] <xemet> I will ask to my friend what he need
[18:39:41] <xemet> He asked me if it was possible to have only two axes
[18:39:51] <xemet> so I think he needs only two axes...
[18:40:13] <xemet> and possibly he has a special machine dedicated to certain tasks where only X and A are needed
[18:40:39] <xemet> so I hope that there are no Y or Z in the G-code he wants to use...
[18:40:43] <xemet> I will ask to him
[18:41:42] <alex_joni> xemet: I think you can also use this: http://axis.unpy.net/01162326817
[18:42:06] <alex_joni> that's gantrykins (written by jepler), but you can use that instead of writing a new (very simple trivkins)
[18:42:33] <xemet> the link doesn't work...
[18:43:02] <alex_joni> works here
[18:43:04] <xemet> so, where could I read the trivial kinematics code...for example the one for XYZA
[18:43:08] <jlmjvm> it woked for me
[18:43:19] <xemet> ok
[18:43:23] <xemet> now it works
[18:43:25] <jlmjvm> cool
[18:43:37] <alex_joni> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/emc/kinematics/trivkins.c?rev=1.12;content-type=text%2Fplain
[18:44:48] <xemet> great...so, tell me, since my problem is very simple, I could simply modify this...
[18:44:52] <xemet> or not?
[18:44:57] <alex_joni> xemet: right
[18:45:00] <xemet> ok
[18:45:26] <alex_joni> xemet: apt-get source emc2
[18:45:32] <alex_joni> sudo apt-get build-dep emc2
[18:45:33] <xemet> only a thing I don't understand very well...
[18:45:40] <alex_joni> and you should be able to recompile emc2
[18:45:49] <xemet> so when I define a axis as Angular in the ini file
[18:46:01] <alex_joni> xemet: that's only for jogging
[18:46:19] <alex_joni> it usually doesn't make much of a difference if it's ANGULAR or LINEAR
[18:46:42] <alex_joni> xemet: and some MAX_VEL & MAX_ACCEL behave a bit different
[18:46:51] <xemet> if I define Y as angular, it moves exactly like X that is linear
[18:46:56] <alex_joni> right
[18:47:14] <alex_joni> xemet: you won't see much on the AXIS display
[18:47:28] <alex_joni> as it doesn't show changes in A,B,C
[18:47:35] <xemet> if I use xyza
[18:47:39] <xemet> the cone rotates...
[18:47:56] <xemet> when I type G0 A90...
[18:47:58] <alex_joni> ok
[18:47:58] <xemet> for example
[18:48:45] <xemet> if I set only two axes...could I continue to see the cone that rotates when I change A or it is not possible?
[18:49:06] <alex_joni> if you change the kinematics it should work
[18:50:18] <xemet> ah ok...
[18:50:39] <xemet> uhm...I'm reading the code for the trivial kinematics...
[18:51:57] <xemet> it seems that the only relevant part for movement is the one with the pos->tran.x = joints[0]; etc..
[18:52:06] <alex_joni> and the reverse one
[18:52:15] <alex_joni> joints[0] = pos->tran.x
[18:52:29] <xemet> let me understand... so if I delete Y and Z
[18:52:55] <xemet> and wite joints[1] = pos -> a
[18:53:04] <xemet> and the reverse
[18:53:13] <alex_joni> pos->a=joints[1]
[18:53:17] <alex_joni> and that'sit
[18:53:40] <xemet> just that?
[18:53:45] <alex_joni> yeap
[18:53:55] <xemet> where is it said that the a axis has to rotate in AXIS?
[18:54:30] <alex_joni> xemet: geometry second course
[18:54:57] <alex_joni> the axes A, B and C are rotating around X, Y and Z
[18:54:58] <xemet> ok
[18:55:10] <xemet> yes, I know
[18:55:49] <xemet> we don't call them ABC here...but I see it is the same
[18:58:09] <alex_joni> usually I call them alpha, beta and gamma
[18:58:35] <xemet> RST here...but I think they could be called how everyone prefers
[18:58:58] <alex_joni> RST are something the emc terminology uses too
[18:59:06] <alex_joni> but those aren't around XYZ afaik
[18:59:26] <alex_joni> RST are thought of some other (wherever)
[18:59:32] <alex_joni> RST are thought of some other (wherever) rotary axes
[18:59:43] <xemet> ok understand
[18:59:58] <xemet> well
[19:00:31] <xemet> so the most difficult thing now is learn to recompile emc2
[19:00:58] <xemet> I compiled it only one time but it was in a directory and I used run in place
[19:01:32] <alex_joni> xemet: that's probably the best thing to do :)
[19:01:38] <alex_joni> or do it like I said above
[19:01:42] <alex_joni> apt-get source emc2
[19:01:44] <xemet> ok
[19:01:47] <alex_joni> sudo apt-get build-dep emc2
[19:01:51] <alex_joni> cd emc2/src
[19:01:52] <xemet> ok
[19:01:57] <alex_joni> make && sudo make install
[19:01:56] <xemet> perfect
[19:02:09] <alex_joni> err.. ./configure --prefix=/usr
[19:02:18] <xemet> ok
[19:02:32] <alex_joni> xemet: or you can use debuild to build a new package (with your modification in it)
[19:02:40] <alex_joni> it will do all is needed for you
[19:02:54] <alex_joni> cd emc2/
[19:03:06] <alex_joni> debuild (you probably need to install with apt-get install debuild)
[19:04:02] <xemet> so, with make && make install
[19:04:16] <alex_joni> sudo make install
[19:04:16] <xemet> it will be installed, not run in place
[19:04:27] <xemet> yes
[19:04:32] <alex_joni> a normal user can't write to places where emc2 needs to be installed
[19:04:40] <alex_joni> also note that there's no easy way to uninstall
[19:04:44] <xemet> yes I've learnt it
[19:04:57] <alex_joni> you need to manually remove all files
[19:05:05] <xemet> no from sinaptyc?
[19:05:08] <alex_joni> the deb method is a bit cleaner/easier
[19:05:24] <alex_joni> xemet: sinaptic doesn't know you copied some files
[19:05:30] <xemet> ah ok
[19:05:30] <alex_joni> (which is what make install does)
[19:05:32] <xemet> but right now
[19:05:46] <xemet> If I don't recompile, I could unistal it from synaptic
[19:05:49] <alex_joni> xemet: you can do another thing
[19:05:49] <xemet> right?
[19:05:52] <alex_joni> right
[19:06:01] <alex_joni> xemet: configure emc2 to work run-in-place
[19:06:05] <alex_joni> and compile it there
[19:06:22] <alex_joni> once you are happy with it, copy the trivkins.ko over the one already installed by the package
[19:06:31] <alex_joni> all other files you won't change anyways
[19:06:45] <xemet> ok understood...it was the first thing I thought...
[19:06:49] <alex_joni> dpkg -L emc2 will give you a list of all files
[19:06:59] <alex_joni> so you can easily find trivkins.ko
[19:07:22] <xemet> I can change the name of the trivkins? Or I will lost the trivial kinematics forever?
[19:07:52] <alex_joni> xemet: you can change the name, but then you won't be able to easily start it (not without hacking emc_module_helper)
[19:08:05] <alex_joni> xemet: you can 'hijack' one of the other kinematics modules
[19:08:25] <xemet> uhmmm understood...so adding a new module is not that easy
[19:08:34] <alex_joni> xemet: call it tripodkins, and use in your hal file (loadrt tripodkins)
[19:09:02] <alex_joni> xemet: we don't run emc2 as root anymore, for that to work there's a small program which does the root things (insmod/rmmod) for us
[19:09:07] <xemet> ok, I will overwrite an existing file
[19:09:19] <alex_joni> but to keep things secure it only does so with filenames it knows about
[19:09:29] <alex_joni> so users can't use that to hack an emc2 machine :)
[19:09:32] <xemet> understand
[19:09:49] <alex_joni> I'm glad you do
[19:10:00] <xemet> many many many thanks!!
[19:10:07] <alex_joni> sure, no problem
[19:10:08] <xemet> for your help
[19:10:09] <tomp> alex_joni: video wall from panasonic http://www.youtube.com/watch?v=s-sx02eiysg "you can play the basket ball" : do 6 of these make a holodeck? xemet: they use you control near the end ;)
[19:12:04] <xemet> just for curiosity...if I would to ADD a new module...what should I change? the emc_mocule_helper?
[19:12:10] <alex_joni> xemet: yes
[19:12:28] <alex_joni> I think I told skunkworks how to do that a while ago
[19:12:33] <alex_joni> and it should be logged by logger_emc
[19:12:36] <jepler> at least in emc 2.1 and later you don't have to change emc_module_helper to add a new module
[19:12:55] <xemet> really?
[19:13:15] <xemet> so I can change the name to trivkins.c and compile it with a new name and add it?
[19:13:43] <alex_joni> jepler: it uses the folder in the whitelist?
[19:13:48] <jepler> alex_joni: yes, basically
[19:14:07] <xemet> so what should I do?
[19:14:12] <jepler> xemet: yes. you must also change the path passed to hal_init()
[19:14:37] <jepler> then compile and install it with: comp --install mykins.c
[19:14:44] <xemet> 1) modify the trivkins.c rename it to for example twokns.c
[19:15:01] <xemet> ah, with comp??
[19:15:12] <xemet> no make...make install etc.?
[19:15:25] <jepler> comp --install does a compile and an install
[19:15:43] <xemet> perfect...
[19:15:49] <xemet> yes I've used it yesterday...
[19:16:01] <jepler> you can also do it by adding lines to the emc2/src/Makefile -- if you prefer to do that, I can tell you how
[19:16:29] <xemet> uhm...if comp works I think it is easier for me...
[19:17:30] <xemet> so just modify trivkins, rename and comp --install
[19:17:34] <xemet> great
[19:17:46] <xemet> it seems very easy
[19:17:50] <xemet> I will try
[19:17:52] <alex_joni> it should be
[19:18:19] <alex_joni> xemet: also, it's good to ask.. there are new things every day :)
[19:18:19] <xemet> fabiotv: ci sei?
[19:18:44] <xemet> this the beautiful thing about emc2...
[19:18:56] <fabiotv> yes cercavo di capire
[19:19:13] <xemet> fabio: forse ho risolto, stasera ti faccio sapere.
[19:19:42] <alex_joni> * alex_joni passes a big wooden bat to xemet
[19:19:54] <alex_joni> xemet: that helps for teaching people
[19:20:04] <xemet> fabio: almeno per ciò che riguarda i due soli assi.
[19:20:21] <fabiotv> si può fare?
[19:20:27] <xemet> Alex please tell me...HOW do you write those purple messages??
[19:20:35] <alex_joni> xemet: purple?
[19:20:41] <xemet> I see...
[19:20:44] <alex_joni> probably because your name is in them D
[19:21:01] <alex_joni> xemet: your client might have nick notifiying or something
[19:21:06] <xemet> * alex_joni passes a big wooden bat to xemet
[19:21:08] <xemet> this
[19:21:21] <alex_joni> I use /me before the command
[19:21:29] <alex_joni> /me does something
[19:21:32] <alex_joni> * alex_joni does something
[19:21:35] <crepincdotcom> * crepincdotcom lolz@y'all
[19:21:38] <Jymmmm> * Jymmmm whacks alex_joni with a clue-by-four!
[19:21:44] <xemet> great
[19:22:14] <xemet> fabiotv: si si può fare
[19:22:34] <xemet> fabiotv: ma prima provo e poi ti faccio sapere e ti do istruzioni
[19:22:44] <fabiotv> ok
[19:22:49] <alex_joni> fabiotv: e molto facile
[19:22:56] <crepincdotcom> why are there no ops in the channel? do you guys just do /ns update whenever you need to kick etc someone?
[19:22:58] <xemet> great
[19:23:21] <xemet> * xemet alex can talk italian. Make an applause to him
[19:23:30] <alex_joni> crepincdotcom: you men /cs ?
[19:23:32] <alex_joni> crepincdotcom: you mean /cs ?
[19:23:43] <crepincdotcom> no i meant ns.... services is all one program
[19:23:50] <crepincdotcom> ns update tells chanserv to update your modes
[19:23:52] <crepincdotcom> ah ok
[19:23:57] <Jymmmm> * Jymmmm just does F12 =)
[19:24:01] <crepincdotcom> lol
[19:24:15] <Jymmmm> or /k alex_joni works too
[19:24:26] <Jymmmm> Here, let me show you...
[19:24:47] <crepincdotcom> * crepincdotcom is is an oper at a small network
[19:24:52] <Jymmmm> Ah, my mistake.... it's /k al<tab><enter>
[19:24:58] <crepincdotcom> lol
[19:25:10] <alex_joni> ns is usually NickServ
[19:25:11] <xemet> well...I go out for dinner
[19:25:17] <alex_joni> cs is ChanServ
[19:25:21] <Jymmmm> and ms is memoserv
[19:25:21] <xemet> thank you all for help!
[19:25:43] <crepincdotcom> alex_joni: theyre the same program, so you tell nameserv to update you, and chanserv does the action
[19:25:45] <Jymmmm> alex_joni speak boy, speak! WOOF
[19:25:45] <crepincdotcom> at least with most services
[19:25:48] <fabiotv> thanks thanks thanks
[19:26:13] <xemet> good byeeeee
[19:26:27] <alex_joni> alex_joni has kicked Jymmmm from #emc
[19:26:34] <crepincdotcom> burn
[19:26:40] <alex_joni> Jymmmm: what? no superpowers?
[19:27:10] <Jymmmm> alex_joni what, you removed me?
[19:27:23] <crepincdotcom> * crepincdotcom hides
[19:28:00] <pier_> pier_ is now known as pier
[19:29:28] <alex_joni> Jymmmm: what did you do?
[19:29:37] <fabiotv> You are angers or games to you of power
[19:29:53] <Jymmmm> alex_joni ?
[19:29:54] <alex_joni> fabiotv: say that in italian
[19:30:01] <alex_joni> 21:23 -!- #emc Cannot send to channel
[19:30:12] <Jymmmm> alex_joni muted you.
[19:30:22] <alex_joni> Jymmmm: that's not nice
[19:30:34] <Jymmmm> alex_joni what, you have ops, whats the problem?
[19:30:53] <alex_joni> Jymmmm: no problem
[19:31:00] <crepincdotcom> +m-v ftww
[19:31:06] <Jymmmm> alex_joni you took away my ops, so I couldn't remove it
[19:32:09] <alex_joni> I did not P
[19:32:16] <Jymmmm> alex_joni liar
[19:33:34] <crepincdotcom> thank you
[19:33:43] <Jymmmm> * Jymmmm sighs.... noobie ops, eeeesh
[19:34:01] <pier> Hi all, yesterday I missed your comments about emc and hot wire foam cutting machine... is there a way to retrieve them?
[19:34:02] <crepincdotcom> he kinda owned himself on that one
[19:34:33] <Jymmmm> crepincdotcom and everyone else too =)
[19:34:42] <crepincdotcom> ;-) no kidding
[19:34:43] <anonimasu> hehe
[19:34:48] <alex_joni> hmm, Jymmmm might be too dangerous for ops in #emc
[19:34:50] <anonimasu> logger_emc: bookmark
[19:34:50] <anonimasu> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-02-23.txt
[19:35:01] <anonimasu> pier: I hope that helps
[19:35:02] <crepincdotcom> this ircd wouldn't even let me /nick bitch-take-off-plus-m
[19:36:06] <Jymmmm> alex_joni theres nothign there of hot wire
[19:36:54] <pier> anonimasu: thanks!
[19:37:01] <Jymmmm> 22
[19:37:20] <Jymmmm> pier which dos program are you using?
[19:37:59] <pier> nope
[19:38:06] <Jymmmm> ?!
[19:38:09] <pier> Jymmmm: sorry
[19:38:38] <Jymmmm> pier and do you have any pics of the final results?
[19:38:38] <pier> I didn't find the reference in that link
[19:38:54] <pier> Jymmmm: I wrote the code myself
[19:39:07] <Jymmmm> pier ah, ok.
[19:39:19] <pier> Jymmmm: cr.. code but planes are flying so...
[19:40:33] <pier> Jymmmm: I am still stuck with dos on account of that program and haven't had time to port it to Linux
[19:40:38] <Jymmmm> pier Gotcha =) I'm gather parts to make one too.
[19:40:57] <Jymmmm> Gawd knows I have enough hot wire
[19:41:58] <pier> Jymmmm: http://pier.unirc.eu/images/foam_cut2.jpg
[19:42:17] <Jymmmm> pier is that tapered on the end?
[19:42:24] <pier> yes
[19:42:35] <pier> it's a zagi reflex
[19:42:39] <Jymmmm> pier how did you account for that?
[19:42:42] <pier> for a tailless
[19:42:43] <Jymmmm> the taper I mean
[19:43:02] <pier> sorry..
[19:43:38] <Jymmmm> pier how did you design the taper, then how did you cut it?
[19:43:40] <Jymmmm> cnc wise
[19:43:49] <pier> I start with different position of the wire end
[19:44:06] <pier> ends
[19:44:23] <pier> one is ahead
[19:44:32] <pier> the right amount
[19:44:44] <Jymmmm> so, it has to make pultiple passes to make a clean taper?
[19:44:49] <Jymmmm> multiple
[19:45:02] <pier> no... one cut
[19:45:16] <pier> the wire starts to move
[19:45:33] <Jymmmm> ok, so you have 4 drivers, one per motor instead of parallel two motors?
[19:45:33] <pier> from the trailing edge
[19:45:40] <pier> yes
[19:45:42] <Jymmmm> k
[19:46:29] <pier> I wrote the first program for a siemens microcontroller in plain asm
[19:46:32] <Jymmmm> I guess it's a lil hard to image w/o seeting it.
[19:46:36] <pier> but I was getting mad
[19:46:42] <Jymmmm> pier you sick puppy you =)
[19:47:32] <Jymmmm> in respect to the ASM that is
[19:47:50] <pier> then when it came to heat compensation I discovered that C and a parallel port would save me from insanity
[19:48:54] <Jymmmm> pier I know there a proggy out there that does foam cutting, but havent' looked at it.
[19:49:07] <Jymmmm> but windows
[19:49:37] <pier> yes I know but I am allergic to Windows
[19:49:46] <Jymmmm> lol, fair enough.
[19:50:03] <pier> an idea dawned to use emc for that purpose
[19:50:25] <Jymmmm> alex_joni should have some ideas on that one
[19:50:46] <pier> but yesterday when I asked the list I had to scoot
[19:56:34] <Jymmmm> I don't see any mention in the logs after you sked initially.
[19:57:00] <pier> I found it
[19:57:02] <Jymmmm> I already have the NiChrome, and just got some slides I can use.
[19:57:13] <Jymmmm> pier where?
[19:57:24] <pier> http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-02-22.txt
[19:57:40] <Jymmmm> pier time?
[19:58:54] <pier> 15:49:04
[20:02:17] <Jymmmm> pier If you get it working, let me know.
[20:04:06] <pier> Jymmmm: it would be nice... yet I am still through the hal tutorial
[20:04:06] <Jymmmm> pier what are you using to draw up plans?
[20:04:19] <pier> profiles?
[20:04:26] <Jymmmm> uh, ok =)
[20:04:37] <alex_joni> pier: you remember to take notes on things that are not accurate in the manual, right?
[20:05:06] <Jymmmm> * Jymmmm is still having a hard time envisioning a wire cutting a smooth taper , even if by hand.
[20:05:08] <pier> mmm not able to make out errors...
[20:05:42] <pier> In the code I wrote you insert the 4 NACA digits
[20:06:01] <pier> to get the profile
[20:06:16] <alex_joni> pier: the HAL tutorial I meant
[20:06:18] <pier> one for the root and one for the tip
[20:06:21] <Jymmmm> pier but what are you using to draw the profile?
[20:06:29] <alex_joni> pier: there might be commands which aren't actual any more
[20:06:31] <Jymmmm> with the taper I mean.
[20:06:33] <pier> I do not draw them
[20:06:39] <Jymmmm> oh
[20:07:16] <pier> it's the program that either make the calculations according to NACA formulae
[20:07:35] <pier> or grab a data profile file
[20:07:51] <pier> You put in the sweep
[20:07:56] <Jymmmm> Well, I'm more interested in drawing/designing such things.
[20:08:03] <pier> the root chord and the tip chord
[20:08:19] <Jymmmm> not necessarily for RC planes, but the taper part for other applications.
[20:08:33] <pier> Jymmmm: it export dxf that you can insert in your drawing
[20:08:50] <Jymmmm> k
[20:08:51] <pier> Jymmmm: ha
[20:10:29] <pier> It would be necessary to provide the Gcode so that emc moves x-y and u-v along different paths
[20:11:15] <Jymmmm> If I could draw somethign 3D, then have it cutout in emc, that would be very cool.
[20:11:15] <pier> alex_joni: ok thanks
[20:11:42] <pier> Jymmmm: that would be great
[20:12:20] <Jymmmm> I also have some 1/2" 10TPI ACME rod already too, so I'm mostly there - parts wise.
[20:12:53] <pier> my machine is a real rattlebone
[20:13:23] <pier> Jymmmm: http://pier.unirc.eu/images/maccali.jpg
[20:13:31] <Jymmmm> heh, I was looking at some designs, scarry. But I though for the framing, I could use EMT (maybe)
[20:14:17] <Jymmmm> pier hey, it's all good, whatever works =)
[20:14:46] <Jymmmm> I just have the sticks of acme rod, as I got them cheap and on sale.
[20:15:03] <pier> Jymmmm: ;)
[20:15:44] <pier> Jymmmm: it's a French design
[20:15:46] <Jymmmm> lol
[20:15:58] <pier> Jymmmm: I mean it
[20:16:02] <Jymmmm> Yeah, I saw that website too =)
[20:18:45] <pier> Yet with emc a sw that corrects the profile with the right amount of offset would be still necessary
[20:19:15] <pier> because of the different cutting speed from root to tip
[20:19:58] <pier> king of tool correction
[20:20:47] <pier> which would be small for the root and greater for the tip tapered profile
[20:25:29] <alex_joni> did you guys look at Joost?
[20:25:49] <alex_joni> https://www.joost.com/about.html
[20:35:54] <alex_joni> unfortunately no linux support
[21:05:22] <Jymmmm> alex_joni: There are many of those types of sites around. Many just have a library of TV shows. But if you really want something interesting, check out TED... http://www.ted.nu/
[21:31:12] <skunkworks> Jymmmm: I have a vp1020a card
[21:36:18] <jepler> http://jet.ro/dismount/ <-- this is a lot of fun
[21:45:56] <skunkworks> jepler: it is actually pretty relaxing. Fun
[22:38:24] <pier> Nite all. Have a nice week end.
[23:33:40] <skunkworks> Jepler: you around?
[23:38:10] <skunkworks> Jepler: I think I figured out the problem with my portable. I was using a 2 ft parallel cable. It is actually from an old zip drive. It worked on the computer I played with at work but it doesn't seem to work with my portable. The second I removed the parallel cable - the encoder counting worked correctly.
[23:38:50] <skunkworks> jepler: so it was me :)
[23:40:03] <skunkworks> (I just can't have my portable power plugged in and the pluto at the same time) I might try to find a different cable and see if it works better
[23:46:33] <jepler> skunkworks: that is interesting
[23:46:47] <jepler> skunkworks: I'm on my way out, but thanks for letting me know
[23:46:54] <jepler> it would be nice to blame it all on cabling :-P
[23:47:59] <CIA-6> 03jepler 07TRUNK * 10emc2/src/hal/user_comps/devices/hal_joystick.c: the spaceball has 6 axes
[23:48:33] <skunkworks> so far so good :)
[23:49:40] <jepler> skunkworks: so you plugged the pluto directly into the laptop's parallel port, no cable?
[23:49:49] <skunkworks> jepler: correct
[23:50:41] <skunkworks> I figured it was short enough not to matter - but I guess I was wrong. Will have to buy an ieee mumble mumble cable. see if that works better
[23:52:40] <skunkworks> I hooked it back in just to be sure - and it didn;t work correctly again.
[23:54:24] <skunkworks> jepler: sorry to worry you.
[23:55:24] <skunkworks> jepler: have fun tonight.