#emc-devel | Logs for 2010-11-15

[00:00:04] <andypugh> The c-code looks OK, it has the pin creation code and the docs.
[00:00:22] <andypugh> So the pin declararation code seems to be OK
[00:01:58] <micges> have you looked around 271 line in c file to locate code that has -> in it?
[00:03:21] <micges> good night all
[00:24:26] <cradek> andypugh: missing } on line 318
[00:25:19] <andypugh> Ah! Thanks.
[00:25:58] <cradek> sure can be hard to find those!
[00:26:30] <andypugh> Especially as the errors were being raised _before_ that point.
[00:26:53] <cradek> yes I don't get that
[00:27:03] <andypugh> It was at line 504 in the .c
[00:27:41] <andypugh> And, seems to be a different issue, as it still gives the same error.
[00:27:42] <cradek> hm, that doesn't fix it
[00:27:48] <cradek> **sigh**
[00:27:55] <andypugh> Well, it needed fixing anyway.
[00:27:58] <andypugh> So thanks.
[00:28:02] <cradek> sure
[00:33:09] <cradek> andypugh: missing ) on 313
[00:33:21] <andypugh> Just fixed it :-)
[00:33:32] <andypugh> and } on 325 too.
[00:33:45] <andypugh> 329..
[00:34:10] <andypugh> copy-paste coding is a good way to multiply errors
[00:36:36] <cradek> 375
[00:37:03] <andypugh> unknown variable?
[00:37:16] <cradek> no, extra } on 375
[00:37:47] <andypugh> 375 or 325?
[00:38:04] <cradek> it's 375 here
[00:38:13] <cradek> right after the if (theta < 0) block
[00:38:30] <cradek> also I bet you are missing break; before case 2:
[00:38:34] <andypugh> I must have already got it, and forgotten
[00:38:46] <andypugh> Yes, the breaks are in too, now.
[00:41:54] <andypugh> I could understand if I had header files with unterminated lines or something, but there isn't.
[01:06:53] <andypugh> cradek: Got it. I had "state" as a pin and a variable. (I changed my mind). The preprocessed c code was defining stat = inst->state and so everything was getting messed up.
[01:08:12] <cradek> good grief
[01:08:24] <cradek> glad you found it.
[01:08:48] <andypugh> I was commenting out huge blocks of code till the problem disappeared
[01:09:16] <cradek> that's a good strategy
[01:09:21] <cradek> binary search
[01:09:24] <andypugh> The problem genuinely was where the compiler said, but there was a macro substitution going on...
[01:09:47] <andypugh> but a lower-case one.
[01:10:11] <cradek> oh, it's obvious now that you say it. dang.
[01:27:10] <SWPLinux> jepler: http://motionographer.com/2008/03/31/larry-cuba-star-wars-computer-animation/
[01:27:40] <cradek> SWPLinux: what's going on at the shop by now?
[01:34:09] <andypugh> Still no dice here, it's another macro substitution problem I feel sure
[01:34:35] <andypugh> hal/components/bldc.comp:304: error: expected ‘{’ before ‘(’ token
[01:34:35] <andypugh> 304 FUNCTION(_) {
[01:35:46] <jepler> you have something called "state", but comp is using that internally as a parameter name in the _ function
[01:35:49] <jepler> call it state_ instead
[01:36:11] <jepler> cradek: we just got back from dinner
[01:36:42] <andypugh> Ah, I was wondering if it was that. But then I tracked down a different problem with "state" (it was a pin _and_ a variable)
[01:36:56] <andypugh> it "type" free?
[01:39:11] <andypugh> Part of the problem is that this is mix-and-match code, I just wanted to see if it would compile before I went to bed. That was 2hrs ago.
[01:44:11] <SWPLinux> cradek: we're wtching some funny youtube videos :)
[01:44:15] <SWPLinux> +a
[01:49:50] <qq-> there is an application which load all rtai modules on ubu -linuxcnc when running in virtualbox ?
[04:24:39] <jepler> http://www.youtube.com/watch?v=L4OTtFpZV7U -- emc moving 4 axes on the "american robot" at emc fest
[04:25:07] <jepler> er, except the link seems to be brken
[04:25:38] <jepler> .. working now
[04:35:29] <KimK> Hey, thanks for the video. The axis with mechanical problems, it's not some kind of e-stop brake, is it? What seems to be the trouble with that axis? I hope Stuart can find the missing amp, but if it was missing a year ago, well...
[04:37:46] <KimK> I didn't know that "wrist" was so flexible, that's pretty cool.
[04:47:43] <KimK> Back in a bit, must shut down to adjust 3.3v/5V jumpers on one port of 5i22
[05:36:34] <SWPLinux> KimK: we're pretty sure it's the stepper drive, not mechanical or the brake
[05:36:51] <SWPLinux> the other drive is on its way here, it was at another location (we'll have it tomorrow)
[05:37:01] <SWPLinux> the articulation in the wrist is pretty strange
[05:37:25] <SWPLinux> there are some drawings here: http://www.americanrobot.com/products_merlin_robots.html
[07:13:11] <cradek> another of the robot: http://www.youtube.com/watch?v=ssrf9PtJIzU
[08:37:06] <CIA-2> EMC: 03cmorley 07v2.4_branch * r650d17fbff58 10/src/emc/usr_intf/pncconf/pncconf.py: fix pncconf's erratic spindle speed display
[08:37:06] <CIA-2> EMC: 03cmorley 07v2.4_branch * r73d418e2d86c 10/src/emc/usr_intf/pncconf/pncconf.py: finish fixing pncconf's launcher icon
[08:37:07] <CIA-2> EMC: 03cmorley 07v2.4_branch * r0a9f5cf47ae3 10/src/emc/usr_intf/stepconf/stepconf.py: change stepconf's spindle-rpm-filtered signal name
[08:37:14] <CIA-2> EMC: 03cmorley 07v2.4_branch * r8e25d5262309 10/src/emc/usr_intf/stepconf/stepconf.py: fix stepconf's erratic display of spindle rpm
[12:00:07] <jthornton_> jthornton_ is now known as jthornton
[14:55:03] <alex_joni> CIA-2: yay
[14:55:07] <alex_joni> cradek: yay
[14:56:22] <alex_joni> SWPadnos: not that strange actually
[14:56:39] <alex_joni> I think it should be fairly easy to get going with genserkins
[14:57:27] <alex_joni> cradek: is it this one? http://www.americanrobot.com/images_products/merlin_elevation_view_of_mr6500_work_envelope.jpg
[14:58:21] <cradek> yes I think so
[14:59:06] <cradek> the catch is that when you rotate the "inner arm" the "outer arm" maintains the same orientation with respect to the ground
[14:59:42] <alex_joni> so it works like a paralellogram?
[15:00:15] <cradek> yes, if I understand what you mean
[15:00:47] <alex_joni> ok, you can "fix" that in 2 ways (that I see immediately)
[15:01:05] <alex_joni> 1. have another "virtual" joint which gets driven with the same angle but reversed
[15:01:08] <cradek> you can rotate just the shoulder by running two joints the same speed
[15:01:41] <alex_joni> so when you jog the inner arm the outer arm maintains the same angle relative to the inner arm
[15:02:02] <cradek> yes
[15:02:24] <cradek> that seems like it could be a simple tweak to pumakins
[15:02:25] <alex_joni> or you do some adding in HAL for the outer arm angle = outer arm angle + inner arm angle
[15:02:35] <alex_joni> I wouldn't use pumakins for this
[15:02:36] <cradek> or should I try genserkins?
[15:02:47] <alex_joni> yup, genserkins should be waaaay faster
[15:03:26] <alex_joni> pumakins has some constants compiled in (by design), and you can change their values, but not make them 0 or add others
[15:03:37] <cradek> ok, I will look at it, I might bug you for help later :-)
[15:03:40] <alex_joni> so if it doesn't completely match, you're out of luck
[15:03:46] <cradek> ok
[15:03:46] <alex_joni> hopefully I'll be around :D
[15:04:04] <skunkworks> neat videos!
[15:04:11] <alex_joni> did you determine the home location?
[15:04:33] <cradek> no - but it has limit switches, so I bet we can home from them
[15:04:46] <cradek> it has level bubbles
[15:04:52] <alex_joni> ok.. I suggest to have that working first - before attempting kins
[15:05:00] <cradek> yes - good idea
[15:05:46] <alex_joni> otherwise you'll redo some of the kins work
[19:08:05] <jepler> yay, home
[19:08:14] <skunkworks> jepler: did you have fun?
[19:08:37] <jepler> skunkworks: yes
[19:08:52] <skunkworks> Great :)
[19:09:21] <jepler> i'm glad to have a half day at home before work tomorrow though
[19:10:15] <skunkworks> I bet
[19:11:04] <skunkworks> did you and chris ride togather?
[19:11:20] <jepler> no
[19:12:25] <cradek> I'm still here, working on robot kinematics now
[19:12:40] <cradek> got the robot homing and scaled, with encoder feedback
[19:13:05] <skunkworks> cradek: Cool!
[19:13:18] <skunkworks> does it look like serialkins will work?
[19:13:18] <jepler> i left just as the fun was starting, I guess.
[19:13:28] <SWPLinux> there are a few issues
[19:13:29] <skunkworks> videos!
[19:13:32] <skunkworks> ;)
[19:13:36] <cradek> it started (for me) once I got well
[19:13:36] <SWPLinux> regardless of the kins used
[19:13:47] <skunkworks> * More videos
[19:14:08] <skunkworks> cradek: food?
[19:14:37] <cradek> ?
[19:14:45] <skunkworks> or flu?
[19:14:47] <cradek> oh is that what did me in? I think so, but I'm not sure.
[19:14:55] <cradek> I was over it in 1.5 days so it was probably food
[19:14:58] <cradek> nobody else got it
[19:15:01] <skunkworks> odd
[19:15:59] <CIA-2> EMC: 03cradek 07joints_axes3 * r4684358308da 10/configs/sim/axis.ini: make jogs have the right accel and velocity
[19:16:16] <skunkworks> how is the joint/axis branch doing?
[19:17:11] <cradek> we're about to try it some more, I think...
[19:26:12] <alex_joni> http://juve.ro/blog-files/photography/01289848744/IMG_0889.JPG
[19:27:06] <alex_joni> cradek: let me know if I can help with the kins
[19:27:38] <cradek> thanks - I was just reading the totally useless comments about joint and link definitions :-)
[19:28:18] <alex_joni> I usually set DH params by starting with 2 joints, and go from there
[19:28:22] <alex_joni> add 3, 4, etc
[19:28:37] <alex_joni> that way you can set joint pos (0,0,0) and check world
[19:28:45] <cradek> what are A, ALPHA, and D?
[19:28:47] <alex_joni> and tweak DH params till it goes the right way
[19:29:33] <alex_joni> look at this: http://www.youtube.com/watch?v=rA9tm0gTln8
[19:32:30] <alex_joni> cradek: that video makes it look soo easy
[19:33:05] <alex_joni> however there are some catches :D, when you actually do it, somehow it's not as easy anymore, and to make things worse there are actually more "denavit-Hartenberg" ways of doing it
[20:25:05] <psha> is there any way to wait for hal component that's started somewhere else?
[20:25:12] <psha> not via loadusr
[20:25:22] <psha> i've got it
[20:26:00] <psha> thanks )
[21:02:45] <cradek> alex_joni: do you have a dimensioned drawing of the puma560 model you used?
[21:04:51] <skunkworks> I thought it was this http://www.electronicsam.com/images/puma/pumaarm.jpg
[21:06:32] <cradek> sweet, thanks
[21:07:26] <skunkworks> cradek: did you get a stepper drive in?
[21:49:34] <alex_joni> cradek: the drawing skunkworks linked, and the home pos is the same one as in that pic
[21:49:45] <cradek> thanks, that was very helpful
[21:50:10] <cradek> I have my home position with the arm straight out - I think that makes alpha3 = 0 but (aside from lengths) everything else is the same
[21:51:26] <cradek> oh also - my robot is right handed - the picture is left - I think that makes d1 positive
[21:52:48] <skunkworks> cradek: you know you need to make a video of it drawing the 'emc axis' splash screen...
[21:52:51] <skunkworks> ;)
[21:53:06] <cradek> I'm trying my best...
[21:53:27] <skunkworks> no pressure ;)
[22:14:06] <alex_joni> cradek: just remember to test a couple joints first
[22:14:21] <alex_joni> you can easily spot sign/direction swaps
[22:33:15] <JT-Hardinge> cradek: robot is looking good
[22:33:26] <JT-Hardinge> you guys sure are camera shy :)
[22:54:10] <CIA-2> EMC: 03seb 07v2.4_branch * rd4e98fa82623 10/tests/hm2-idrom/skip: skip the hm2_idrom test on 64-bit hardy
[23:04:07] <KimK> cradek: Seeing that it's 5pm, how many EMC'ers are still hanging in with you there in Wichita?
[23:07:47] <alex_joni> g'night all
[23:08:15] <KimK> goodnight
[23:19:13] <skunkworks> logger_dev: bookmark
[23:19:13] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2010-11-15.txt