#emc-devel | Logs for 2010-08-13

[00:15:51] <skunkworks> hmm - I don't know if calculating acutaly commanded rpm is going to be stable enough. when you take (motion.spindle-speed-out)/(halui.spindle-override.value) it doesn't stay constant as you slide the slider. I don't know if it is an issue. if it should be 3000 - I can see it swing 2500-3500 atleast for an instant.
[00:16:17] <skunkworks> As I slide the override slider.
[00:17:07] <skunkworks> I suppose there is a delay between when the FO slider is read and the commanded speed is changed.
[00:17:23] <skunkworks> one lags or leads the other.
[00:19:41] <skunkworks> although I would only be using that during spindle on - to see what actual rpm emc is commanding.. hmm. might have to play with it some more.
[00:55:38] <skunkworks> so it comes back to - it would be nice to have the commanded rpm a pin :) (unmolested)
[00:56:37] <skunkworks> not that I can't make this work. The jury is still out.
[01:58:23] <jepler> skunkworks: you're welcome to test this .. if it works I can push it to master. http://emergent.unpy.net/files/sandbox/0001-Put-commanded-spindle-speed-on-a-hal-pin.patch
[02:03:22] <dgarr> jepler: version2: http://www.panix.com/~dgarrett/stuff/mdi_sub_prob_v2.txt
[02:06:22] <jepler> + status = execute(); // special handling for mdi errors
[02:06:23] <jepler> + if (status != INTERP_OK) {
[02:06:27] <jepler> ah yes, this looks important
[02:07:39] <jepler> you store the error because reset() would otherwise eat it?
[02:09:29] <dgarr> i'm not sure, the error doesn't get passed up somehow in mdi, i just type and test until the result seems right
[02:14:23] <dgarr> and it is retrieving the error (not storing) there
[02:17:25] <jepler> OK, let me see what I can make of it
[02:18:47] <jepler> well, it tests good
[02:20:06] <jepler> argh, check that. I found a different, more convoluted way of making it unhappy :(
[02:20:32] <jepler> main.ngc calls foo.ngc which calls good.ngc and then bad.ngc
[02:21:00] <jepler> on the first run it prints "Unable to open file <bad>" as desired, but on the second run it prints "File ended with no percent sign or program end"
[02:22:30] <jepler> (main.ngc is loaded and run)
[02:24:06] <jepler> though that's no *worse* than current v2.4_branch, and a reload gets it back into a sane state
[02:35:04] <jepler> dgarr: thanks, I pushed a slightly tweaked version
[02:35:38] <CIA-5> EMC: 03jepler 07v2.4_branch * ra38f7094a05d 10/src/emc/rs274ngc/ (interp_o_word.cc rs274ngc_pre.cc rs274ngc_return.hh): Improve handling for mdi invoking nonexistent sub
[02:35:46] <CIA-5> EMC: 03jepler 07v2.4_branch * rc57f2d242e75 10/src/emc/iotask/ioControl.cc: Fix current tool information after reloading table
[02:35:47] <CIA-5> EMC: 03jepler 07v2.4_branch * r99ca8582d14c 10/tcl/bin/emccalib.tcl: clean up debugging print
[03:19:18] <Dave911_> Dave911_ is now known as Dave911
[10:56:11] <skunkworks> jepler: Cool - I will try it today! thanks!
[10:59:30] <alex_joni> skunkworks: feeling better?
[11:12:35] <skunkworks> yes - finally. Thanks alex
[11:12:44] <skunkworks> biab - Have to go to work
[12:32:50] <skunkworks> logger_dev: bookmark
[12:32:50] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2010-08-13.txt
[12:38:31] <alex_joni> didn't miss anything ;)
[12:40:56] <skunkworks> I needed the url :P
[12:41:41] <skunkworks> we have gotten an inch of rain so far in 1 hour
[12:43:21] <alex_joni> http://juve.ro/blog/projects/01281702733
[12:48:03] <skunkworks> that was weird. Cool alex!
[12:50:04] <alex_joni> maybe a month till it's done
[12:52:29] <skunkworks> jepler: wow that seemed a little more than trivial - you had to touch 4 files.
[13:10:54] <dgarr> jepler: thanks for improving the patch, new one for consideration: http://www.panix.com/~dgarrett/stuff/0001-support-multiple-DISPLAY-PROGRAM_PREFIX-directories.patch
[13:12:48] <alex_joni> dgarr: AXIS uses PROGRAM_PREFIX to open ngc files by default, how does this work then?
[13:13:01] <alex_joni> opens the first/last/random PROGRAM_PREFIX entry?
[13:13:29] <alex_joni> s/AXIS/emc2 guis (AXIS, tkemc, mini, keystick,..)/
[13:13:34] <dgarr> i didn't know that
[13:13:56] <dgarr> i suppose it could be arranged so it used the first directory
[13:14:40] <dgarr> do you think it is worth pursuing?
[13:15:00] <skunkworks_> jepler: What am I doing wrong? http://pastebin.ca/1916372
[13:15:37] <alex_joni> skunkworks_: look at axis.py and see why it doesn't apply?
[13:15:54] <alex_joni> dgarr: no idea sorry ;)
[13:15:55] <skunkworks> heh - you're funny.
[13:16:13] <alex_joni> in all seriousness, I've been told that before
[13:16:16] <skunkworks> oh - I remember seb saying that it actually puts info there.
[13:23:21] <dgarr> jepler: maybe ii would be better to make a new ini variable for specifying multiple directories for subroutines, i'll work on that
[13:26:22] <skunkworks> confused. I assumed 3890 was a line number - but the axis.py only has 3187 lines
[13:36:35] <cradek> dgarr: I think it'd be nice to have different directories for storing (main) gcode files vs subroutine files
[13:37:08] <cradek> they even have a different format - for instance it's an error to "open" a file with O-sub and no M2
[13:37:43] <cradek> (touchy does this accidentally, since the interpreter looks at PROGRAM_PREFIX and touchy is hardcoded for ~/emc2/nc_files)
[13:39:10] <dgarr> i think you're right, i'll post a new patch later or maybe tomorrow
[13:39:18] <cradek> cool
[13:40:13] <cradek> they're a lot like macros
[13:46:06] <alex_joni> the only thing missing is aliasing g-codes for procedures
[13:46:24] <alex_joni> something like M6 -> O<toolchange>call
[13:47:16] <cradek> yeah but that's not the only problem with that idea - currently a lot of stuff doesn't work in MDI O-call (I think anything that waits for IO is broken)
[13:47:35] <alex_joni> passing params is also a problem
[13:48:07] <alex_joni> like defining Pxx is param #1, Qxx param #2, etc
[13:54:43] <cradek> I guess I don't care about that very much, but I wish I could do O<probe-left-side> call
[14:14:13] <maximilian_h> I found a bug in the stg driver, could somebody please a put a bugfix in the repository
[14:14:16] <maximilian_h> ?
[14:14:55] <cradek> do you have a patch?
[14:15:19] <maximilian_h> just a oneliner, details are in the email that I've just sent to emc-users
[14:15:56] <maximilian_h> but effects can be quite nasty :(
[14:16:07] <cradek> looking
[14:19:00] <cradek> you had offset=0 and gain=1 and you tested with setp value=-10 through value=+10 and measured with voltmeter?
[14:19:05] <maximilian_h> yes
[14:19:22] <maximilian_h> offset and gain are at 0 and 1
[14:19:36] <maximilian_h> and I measured with a voltmeter to reproduce
[14:19:53] <maximilian_h> the problem is that the value gets inverted
[14:19:54] <cradek> great testing
[14:20:01] <cradek> ok I understand
[14:20:03] <maximilian_h> and 0x0000 inverted is 0xFFFF
[14:20:12] <maximilian_h> which is way out of bounds
[14:20:17] <cradek> strange, volts = (value - offset) * gain
[14:20:22] <maximilian_h> so -10V outputs +10V
[14:20:23] <cradek> I would expect (value * gain) + offset
[14:20:27] <maximilian_h> not good
[14:20:39] <cradek> yes that is very bad!
[14:21:09] <cradek> looks like this is used for both stg1 and stg2
[14:21:12] <cradek> I wonder if they work the same
[14:21:18] <maximilian_h> yes volts = (value - offset) * gain, but you are not writing floats to the register either ;)
[14:21:47] <maximilian_h> yes, stg1 and stg2 cards work the same in that regard. At least the manual is the same
[14:22:20] <cradek> ok, thanks for checking that
[14:22:46] <cradek> do you want to make a git patch for me to apply so you get authorship of this change, or do you want me to make the one line change for you?
[14:22:55] <maximilian_h> both times manual p. 30
[14:23:40] <maximilian_h> make it for me please. I have spent way too much time already on this today
[14:24:16] <cradek> ok, no problem
[14:24:36] <maximilian_h> good, thanks, then it's solved
[14:24:44] <cradek> thanks for the fix! that's an important one.
[14:24:52] <cradek> I will make the change in the 2.4 branch.
[14:25:17] <maximilian_h> no prob, affected me too, so I was quite motivated ;)
[14:25:46] <maximilian_h> 2.3 is affected too
[14:26:07] <maximilian_h> although the code is the same vor 2.3 and 2.4, I've checked
[14:26:08] <cradek> that's true, but I kind of doubt we will make another 2.3 release
[14:26:14] <maximilian_h> true
[14:26:30] <maximilian_h> and I wrote the email to the list. So 2.3 should be notified
[14:27:11] <cradek> yes true
[14:27:32] <maximilian_h> ok, I'll get back to my other work, bye
[14:29:05] <cradek> thanks, bye
[14:33:53] <CIA-5> EMC: 03cradek 07v2.4_branch * r8511f46257f0 10/src/hal/drivers/hal_stg.c: Fix incorrect negative DAC output on STG cards
[14:38:41] <skunkworks> I wonder if I can ignore the axis.py changes.
[14:39:08] <skunkworks> seems unrelated.
[15:21:50] <dgarr> revised patch for multiple subroutine directories: http://www.panix.com/~dgarrett/stuff/0001-new-ini-parameter-DISPLAY-SUBROUTINES.patch
[15:42:55] <maximilian_h> Hey Dgarr, will your patch also work with the user defined M command which reside in PREFIX_PATH too ?
[15:44:10] <dgarr> i don't think so, i will look at that after resolution of this one, thanks for the idea
[15:51:27] <maximilian_h> dgarr: Thanks, that would be really great
[17:03:10] <skunkworks> can I run sim on a 64bit install?
[17:03:17] <skunkworks> of lucid?
[17:19:17] <awallin> sure why not?
[18:45:51] <alex_joni> skunkworks: you just have to compile it
[18:54:58] <skunkworks> I know. But I ran into a little problem with my laptop.
[18:55:32] <skunkworks> I changed the 3 files (not the axis.py) and compiled it - but the opengl is screwed on this computer.
[18:55:58] <skunkworks> so I thought - I have been meaning to install lucid 64 - so that is what I am doing.
[19:03:37] <cradek> dgarr: (if you see this later) I made a few changes and here is a new suggested patch for your review: http://timeguy.com/cradek-files/emc/0001-new-ini-parameter-directory-search-list-for-subrouti.patch
[20:11:46] <skunkworks> wow - lucid64 installed - emc2 compiled and runs.
[20:11:59] <skunkworks> note to self - don't install the closed source video drivers. ever
[20:12:16] <cradek> oops
[20:46:51] <skunkworks> jepler: I added the 3 lines to the 3 files and did a make. It looks to be exactly what I need. Cool
[20:46:54] <skunkworks> thanks again