00:38:58 acemi has quit 01:53:48 roltek has quit 01:55:36 robin_z has joined #emc 01:55:42 meep? 02:04:25 robin_z has left #emc 02:28:40 Telefunken has joined #EMC 02:29:38 Ephexis has quit 02:30:07 question. emc2. Is there an example of a '1 bit per winding' setup on the parallel port? 02:30:45 I only see dir and step in stepgen when I 'halcmd show'. 02:31:31 I need to have outputs like '1000' '1100' '0100' 0110' etc 02:33:46 jepler: I'm sure it's in there (without looking at the source) 02:33:58 jepler: something about phase drive 02:34:12 I remember seeing all the tables for the different stepping sequences 02:34:33 src/hal/components/freqgen.c 02:34:54 /* lookup tables for stepping types 2 and higher */ 02:35:09 yeah that 02:35:25 there are a few to choose from 02:37:25 How do I control the arguments to 'insmod'? 02:37:34 in the hal file 02:37:37 look at the parallel port one 02:38:28 core_stepper.hal:loadrt stepgen step_type=0,0,0 02:38:59 so I want to .. copy core_stepper.hal and edit it? 02:39:18 sure, I guess 02:39:23 copy it into your config dir 02:39:27 change your ini to point to it 02:39:31 Hal_Introduction.pdf has sets of output timing stuff. 02:39:59 somewhere around page 38-46 02:40:33 looks like type 9, from stepgen.c's comments about the tables 02:41:39 03cradek * 10emc2/ (6 files in 4 dirs): (log message trimmed) 02:41:40 Fix up make install target, debian build, make deb target, etc. 02:41:40 To use it, I recommend: 02:41:40 ./configure --prefix=/usr --sysconfdir=/etc && make deb 02:41:40 The default configure settings will put a lot of things in /usr/local, 02:41:40 which I think is somewhat nonstandard for a deb. I can't decide whether 02:41:42 this flexibility (and therefore the requirement that the deb creator 02:45:38 03rayhenry * 10emc2/tcl/bin/halconfig.tcl: Started on watch and other mode stuff. 02:54:32 rayh has quit 02:59:39 haven't hooked up the hardware yet, but I have some plausible looking signals in halscope... 02:59:56 cool 03:00:17 does the path follow the program? 03:00:33 I dunno .. actually it's running tkemc as the front end 03:00:41 and I've just been doing 'g0's 03:01:32 oh good -- halscope seems to remember those things I painfully configured last run 03:01:53 yeah, it's good that way 03:12:46 just don't load it if the signals aren't present, or it'll lose (parts of) that painfully crafted config 03:14:02 I can't find the driver board or my parallel cable 03:14:04 * jepler keeps looking 03:14:20 I sure know how that feels 03:14:27 I was looking for a particular video cable tonight 03:14:31 is it plugged in? ;) 03:18:05 SWPadnos: no, but thanks for asking 03:18:23 heh - always the first tech support question 03:18:38 I feel your pain though - I actually managed to lose an MSC catalog in my office 03:23:55 well, I found a parallel cable, the wrong driver board, and the wrong motors .. but at least I should be able to spin 'em 03:24:21 but the windings must be in a different order than I thought 03:27:05 can you change that in the halfile? 03:27:07 yay, with a little tweak to the .hal file the motors are spinning right 03:27:22 slick 03:27:39 cradek: yes, connecting the winding signals to different parport pins is easy .. even I could figure it out 03:29:47 iirc, the etch-a-sketch was 68 half-steps per inch. With LINEAR_UNITS = 0.039... and INPUT_SCALE = 4000, this lead to weird behavior under emc1 03:30:52 so do I want to try a value like .0032... (feet?) and INPUT_SCALE=816? 03:31:01 nope 03:31:08 nope? 03:31:12 the interpreter won't work with anything but the standard few values 03:31:16 oh heck 03:31:27 er, the first time I think I meant INPUT_SCALE = 68 03:31:44 so what do I do? 03:31:49 allowing an arbitrary user unit in the config is strange, because we expect it to be one of about three values and do strange things otherwise. 03:32:05 but with a small INPUT_SCALE things go to hell 03:32:11 have you verified that it works incorrectly? 03:32:12 or at least they did in emc1 03:32:35 you could configure it the way you think is right, then find the bug if there is one... 03:33:08 ooh. axis bug 03:33:21 warnings.append(_("Program exceeds machine maximum on axis %s") % axisnames[i]) 03:37:44 I don't spot it 03:37:51 what's wrong with it? 03:38:44 the loop above goes to 3 all the time 03:39:05 I'm getting joint following errors 03:39:39 I wonder if you're getting a wild path 03:40:06 the backplot looks sane, but I think that's the commanded position, not actual? 03:40:15 now I'm getting 'can't do that (EMC_TASK_PLAN_RUN) in manual mode' 03:40:29 that's the bug I got too 03:40:35 you said there was a fix or workaround? 03:41:37 patch in /query 03:41:50 found it in the bug report too 03:41:51 buried in prose 03:41:55 how wrong do you think the fix is? 03:42:02 with an obsolete line number 03:42:08 no idea 03:42:16 jmk would know 03:42:46 it "seems to work" 03:42:56 but obviously the bug is somewhere else 03:43:38 the bug is not new 03:43:46 I don't remember you mentioning it before, though. 03:43:49 it's new to me 03:44:00 rtai or my faster machine or something must trigger it 03:44:14 I've never used rtai before a few days ago 03:45:45 something's really wrong now 03:46:00 but I'd done an 'update' earlier and not rebuilt, so who knows if it was the change I just made, or something else from earlier 03:46:37 axis comes up without estop pressed, and F1, F2 does nothing 03:46:43 alt-f4 takes a few seconds to do anything 03:46:46 no errors on the terminal 03:46:58 try make clean (cough cough) 03:49:26 I guess if each step is 1/68, then MIN_FERROR had better be bigger than 1/68 03:49:42 that's sure true 03:51:50 maybe it's right 03:52:08 but I was only going 12 ips, not 120ips 03:52:27 ipm? 03:52:31 er yes 03:52:31 ipm 03:52:37 well, I commanded f12 03:52:45 I have no idea what these numbers in the .ini mean 03:53:02 I changed MIN_FERROR = .05 (approx. 4 steps) 03:53:14 now if I think I can do 120ipm how do I set my MAX_VELOCITY ? 03:53:25 2 03:53:31 user units per sec 03:54:04 the same in in [TRAJ] and [AXIS_i] 03:54:12 sure 03:54:28 following error again 03:54:39 axis or joint? 03:54:44 joint 0 03:54:49 doing a move along x 03:54:55 did you set headroom for stepgen? 03:55:53 MAX_VELOCITY = 1.2 03:55:56 for accel? No. I didn't change any of the accel numbers from what was in the config I started with 03:55:57 STEPGEN_MAXVEL = 1.4 03:56:10 oh 03:56:13 forget it 03:56:18 be sure the STEPGENs are higher 03:56:27 for vel and accel both 03:56:48 I got a following error again, but after running part of the program at a very high speed 03:57:02 I used MAX_VELOCITY = 2, STEPGEN_MAXVEL=2.2 03:57:23 that seems good enough 03:57:26 same with accel? 03:57:35 your accel should probably be MUCH higher than the default 03:57:54 well it's 20 by default I guess 03:57:56 maybe that's fine 03:57:56 It gets the following error the first time the velocity goes from +max to -max on an axis 03:58:06 on a jog? 03:58:09 no, on a program 03:58:11 jogs are painfully slow 03:58:40 maybe you should open up your STEPGEN and FERROR settings and see what it does 03:58:44 you may be getting strange paths 04:00:02 I made a big increase in STEPGEN_MAXACCEL from 21 to 41 04:00:11 now I don't get following errors 04:00:38 the backplot looks good, except that all the corners are quite rounded .. 04:02:35 I know that with emc1, the axis backplot showed things all screwed up 04:02:46 now the backplot looks quite correct 04:02:55 I wish I had the right setup to actually connect this to the etch-a-sketch 04:04:04 when I do, the next problem will be the backlash 04:07:08 come to think of it, I haven't seen the etch-a-sketch either! 04:07:21 goodnight guys ... it's exciting to finally run emc2! 04:16:19 slick 04:16:28 I'm glad to hear it works better than emc1! 04:23:18 http://emergent.unpy.net/index.cgi-files/sandbox/axis-etch-gasket.png 04:23:21 http://emergent.unpy.net/index.cgi-files/sandbox/axis-etch-gasket2.png 04:26:59 what feedrate are you running that at? 04:30:20 SWPadnos: 120 inches / minute 04:30:25 SWPadnos: accel 20 04:30:39 OK. and what's the TASK_CYCLE time? 04:31:13 most things are unchanged from the default .. I copied the stepper config directory to start with 04:31:35 does AXIS read the CYCLE_TIME from the ini? 04:31:49 (in the DISPLAY section) 04:31:54 I found the etch-a-sketch and the right driver board .. it works, looks like my custom software did before I added backlash compensation. 04:31:57 SWPadnos: no, it ignores it. 04:32:05 jepler: cool!! 04:32:14 cool 04:32:39 what's the update rate for axis? 04:32:41 SWPadnos: sorry, I don't mean to be evasive about the values ... wireless doesn't work for me with the realtime kernel, so I'm on a different computer right now. 04:32:45 if the value is important, I can go look 04:32:53 no problem 04:33:09 I have the default stepper_inch up right now, so I can look for myself 04:33:28 looks like 20ms 04:33:30 in axis 04:33:35 that's prety fast 04:33:38 pretty 04:33:56 I'm pleasantly surprised that AXIS nicely handles a 2-axis machine 04:34:13 I haven't seen it running that way before 04:34:38 jepler: AXIS on my accelerated nvidia card looks just great 04:34:39 well - those tiny triangles are 5/32 inch across, so at 120 IPM it's only 78 ms per edge 04:35:24 cradek: tried turning on __GL_FSAA_MODE? 04:35:31 heh 04:35:35 nope 04:35:39 not until ... now 04:35:53 if only I could run emc on a *real* meachine 04:35:55 what do I set it to? 04:36:03 16x, of course 04:36:21 cradek: Try a small integer -- 1, 2, 3 04:36:23 (what that's called in GL, I have no idea) 04:36:31 the valid numbers and their meaning differ from card to card 04:36:55 SWPadnos: The screenshot wasn't really intended to question emc2's blending .. I think it's probably doing a fine job 04:37:07 SWPadnos: if you interpreted it that way, it wasn't my intent. 04:37:18 no problem - I'm just curious about the rounding 04:37:27 it's the kind of question that a new user would ask 04:37:35 now it looks fuzzy 04:37:48 cradek: that's what fsaa is supposed to do... 04:38:13 if you were using some texture for the lines, it might look better 04:38:24 but probably not ;) 04:38:40 texture for the lines?? What are you smoking? 04:38:46 heh 04:38:59 make the lines like the 3d_pipes screensaver 04:39:11 jepler: the small integers all make it look differently-crappy 04:39:23 cradek: forget I said anything 04:39:36 can I forget SWP said anything too? 04:39:40 cradek: I found that for day-job-application it was a general improvement, but only after tweaking xgamma too 04:39:41 feel free 04:40:07 it's nice for rendered scenes, but not for CAD / technical data 04:40:17 SWPadnos: despite using opengl, axis is intended to work OK on systems without massive heat-producing 3d accelerators... 04:40:28 and it does 04:40:32 I have a massive heat generator in my big rig 04:40:46 and I usually use that via remote X 04:41:13 glxgears is nice and quick, but not as fast as on the big machine "natively" 04:42:07 goodnight .. for real this time 04:42:12 chris@buster2:~$ glxgears -printfps 04:42:12 13766 frames in 5.0 seconds = 2753.168 FPS 04:42:22 correction - it's slow as hell on the emc machine 04:42:26 and fast on the big machine 04:42:42 125 frames in 5.0 seconds = 25.000 FPS 04:42:44 vs 04:43:02 58483 frames in 5.0 seconds = 11696.417 FPS 04:43:14 it's incredible what you can get with a $40 video card nowadays 04:43:20 slight difference there 04:43:24 yes 04:43:38 if only I could use a video card with the embedded LCD on the emc machine :( 04:44:04 ahh 04:44:09 is it a laptop? 04:44:10 I may end up using a mini-ITX unit instead, though I'd prefer the touchscreen 04:44:28 that I already have 04:44:33 do you use AXIS on a touchscreen? 04:44:47 not yet, but I'll give it a try sometime soon 04:44:50 800x600, too 04:44:54 I think it should work 04:44:58 it fits nicely on 800x600 04:45:05 should - it's just a mouse, but left button only 04:45:17 and no dragging, so all the perspective stuff won't work 04:45:24 (same with the missing middle button) 04:45:28 you can switch between the preset views with the keyboard 04:45:34 yep 04:45:48 there will be a mouse or trackball as well, so that shouldn't be an issue 04:45:48 hmm, no dragging... ouch 04:45:51 yep 04:45:52 ah 04:45:52 Telefunken has quit 04:45:58 it may work, but you get a click first, I think 04:46:10 simple matter of programming 04:46:25 some drivers allow a tap-drag equence 04:57:39 Ephexis has joined #EMC 05:01:21 Ephexis has quit 05:02:12 ccjoe has quit 05:03:30 ccjoe has joined #emc 05:11:34 Ephexis has joined #EMC 06:23:19 SWPadnos_ has quit 09:08:51 Jymmm has quit 09:39:58 Jacky^ has joined #emc 09:40:14 morning 09:52:50 mornin 09:53:09 hi fenn 11:51:13 Jacky^: ov00975.mpg <- why do i have this on my computer?? 11:51:29 er, mov00975.mpg 11:51:48 fenn: because you watch too much pr0n 11:51:55 :D 11:52:31 pr0n.. you wanna se pr0n huh 11:52:33 http://www.theebelinggroup.com/site/videos/nakd/nakd_box.mov 11:54:37 for cardboard lovers 11:56:58 nice video 11:59:23 arg i ought to be waking up now and i'm not even tired 12:01:01 im tryng to get a bit of order in my documents, medical, bank, car, taxes .. 12:01:13 damn papers, it will take to me all day 12:01:29 the 12:02:49 I hate papers 12:03:44 and our minister is sending a book to all families to explain what digital revolution is ! 12:03:51 damn it O_O 12:04:25 digital revolution, like electronic piracy? 12:04:28 ;) 12:04:44 http://www.beppegrillo.it/eng/2006/01/stanca_express.html#comments 12:04:49 read that .. 12:04:49 i read it 12:04:56 damn .. 12:04:59 :/ 12:05:12 beppe's "you must write your full legal name" thing's kinda weird 12:05:29 what's the point of writing your real name to a dissident website? 12:05:30 Ive something like 3000 docs here .. 12:05:49 Beppe is his name 12:06:35 I have to split these docs 12:06:51 I will do a small mountain of paper 12:06:58 then burn it ! 12:07:29 wonder how small it will be 12:07:33 O_O 12:08:01 beppe's comment field says next to it "We remind you that anonymous messages (without real first name and surname) will be cancelled." 12:08:13 oh yes.. 12:08:22 but many peoples use nicks 12:08:32 its not so important 12:08:39 I always used my real name 12:08:57 it would be a trasparent policy of beppe grillo 12:09:15 he like to talk by real name 12:09:21 sometimes when i sign petitions i feel like i'm adding my name to a death list 12:09:28 of course it isnt applicable at all 12:11:06 ahah I found some signature on the petition against software patences by bill g. 12:11:08 lol 12:11:17 want to see something that will make you sick and angry? 12:11:32 lets see.. 12:13:47 so far nobody's taken me up on my offer 12:14:02 :) 12:15:40 uf.. i think the chocolate just wore off.. g'night 12:16:09 im going for the lunch 12:16:13 bye :) 12:16:21 Jacky^ is now known as Jacky^afk 13:10:52 rayh has joined #emc 13:13:16 acemi has joined #emc 13:26:35 Jacky^afk is now known as Jacky^ 13:32:53 5% [46 libx11-dev 407036/1394kB 29%] [29 acroread 2393600/19,9MB 12%] 491kB/s 24m4s 13:33:11 seems a good speed :D 13:36:23 700 mb of updates after 4 weeks .. damn unstable :P 13:36:55 evils not developers 13:38:53 wonder how a user can mantain his distro updated without a fast dsl 13:41:47 who is actually maintaining BDI ? 13:41:53 The only way is to use stable. 13:42:07 rayh: yeah, testing is pretty nice too 13:42:18 maintaining BDI is Paul_c 13:42:33 irts a lot of time i dont see him around 13:42:38 its* 13:42:40 I built a EMC using Knoppix a while back. That is testing 13:42:56 I could not keep up with my dialup. 13:44:05 Im planning to build a Live CD distro for Roboitalia community 13:45:01 Id like to use a RT kernel, such BDI to drive a cnc, and a lot of tools for robotics use 13:45:21 Sure. 13:45:32 rayh: im member of this community http://www.roboitalia.com/ 13:45:43 That is what the vehicle people at NIST do. 13:46:38 skunkworks_wrk has joined #emc 13:46:39 Roboitalia have about 3,376 members now 13:47:03 Neet. 13:47:08 yeah 13:47:52 many of these members don't know nothing about Linux, Emc, .. 13:48:18 someone ask for a software to program microcontrollers in Linux 13:48:45 that why im thinking that a live cd (in italian) could be useful for the community 13:49:17 including some EMC translated doc 13:49:43 You bet it will be a good thing. 13:49:44 BDI shoud be a good point to start 13:50:12 I also look with a great interest to the Puppy distro 13:50:45 I need to get more help and info about Rights terms too 13:53:29 we are actually suffering a lot the chinese competition 13:54:24 Tell me about competition. 13:55:02 I think they take the advantage of read other language, english, italian 13:55:14 where we don't know a word of chinese 13:55:16 :/ 13:55:34 they are very strong 13:55:49 Sure. They catch on quick about competition. 13:56:05 too low cost too 13:56:21 especuially in countries like korea taiwain 13:56:27 taiwan* 13:56:45 Yep. And their quality is getting much better than it used to be. 14:01:11 they are going up in economy too, yesterday I heard china taked the 5 place to the french 14:01:54 by this step, they will dominate the world soon 14:02:32 china get 10 % more in economy in the latest year 14:16:42 Jacky^: ubuntu (and maybe debian) have a complete set of development packages for avr available (gcc and software to program the chips with inexpensive parallel-port devices) 14:17:23 Jacky^: as well as 'sdcc', a C compiler for a number of other devices, including PIC. I haven't used sdcc 14:23:48 Jacky^: simulator packages .. 'gpsim' and 'nitpic' (pic), 'simulavr', 'sdcc-ucsim' (others) 14:23:58 not sure about packages to program anything other than avr, though 14:24:39 Jacky^_ has joined #emc 14:25:02 Jacky^ has quit 14:25:15 Jacky^_ has quit 14:25:23 Jacky^ has joined #emc 14:26:03 jepler: I had to reboot, but I read .. ty 14:26:29 here's something called 'openwince-jtag' 14:26:44 the weak point may be the lack of a nice IDE, which many users want. I don't miss it. 14:26:54 good 14:28:53 which microcontroller are the roboitalia people most interested in? My experience is limited to avr. 14:29:53 I think avr should be ok for must case 14:30:19 I got some good suggestion by LawrenceG too 14:30:22 avr is nice .. in-circuit programming, hardware UART, analog comparator, PWM, external interrupts, and analog comparator on the chip I usually use (at90s2313) 14:30:39 yeah .. 14:31:32 some newbie people is starting using old microcontrollers 16F and other too 14:32:24 Id like to find the way to offer a complete solution for all 14:32:39 may using different apps too, its not a problem 14:32:55 advanced users will use avr 14:34:28 im not familiar with microcontrollers, just want to find a good software answer for theyr needs 14:35:26 of course, lots of user are using windows do p�rogram theyr chip 14:35:46 Id like to find a good Linux alternative 14:35:50 I think that "winavr" is just avr-gcc plus a graphical front end 14:36:35 it should be enough for must case 14:37:55 now tryng to get the community out of the "shell" 14:38:41 starting some wikis and introducing the concept of Free 14:39:20 its hard.. many peoples are closed around theyr ideas, but im tryng 14:39:50 Jacky^: "you can lead a horse to water, but you can't make him drink" 14:39:55 so, I thinked I live cd could be the best way 14:40:17 I'll be back in a few hours. 14:40:26 ok, ty 15:18:50 k4ts has joined #emc 15:20:15 hello 15:22:23 k4ts has quit 16:00:55 Jymmm has joined #emc 16:41:02 skunkworks_wrk has quit 17:29:28 03alex_joni * 10emc2/emc2icon.png: 17:29:29 added Icon for emc2, 17:29:29 thanks to fenn for drawing the little fellow 17:34:12 Jymmm is now known as JymmmAFK 17:52:31 I filed an internet order with Grainger...and they said that they could not identify me as a business and to send a copy of a business liscense 17:52:43 I can do that....but why? 17:53:10 They are not "wholesale"....their prices are retail. 17:54:26 Rather than fax the license copy I have a mind to say "fine...i'll go to MSC." 17:54:30