#emc-devel | Logs for 2008-12-05

Back
[01:30:30] <SWPadnos> BigJohnT, the manual or the drive? (is tiny) :)
[01:59:08] <SWPadnos> oh good - Python 3.0 was released today
[01:59:26] <cradek> so, time to rewrite everything
[01:59:53] <SWPadnos> yes, apparently it's the first "intentionally backwards-incompatible" version
[02:02:20] <LawrenceG> :{{ I'm still learning the first version
[02:02:30] <SWPadnos> second ;)
[02:02:37] <SWPadnos> at least
[02:03:25] <LawrenceG> applying book from 1.5, coding in 2.5,,,, practice for 3 I guess
[02:03:32] <SWPadnos> heh
[02:04:22] <LawrenceG> very nice language.... I wish it would displace java
[02:04:31] <SWPadnos> hmmm. print is now a function - that'll require a lot of changes
[02:05:11] <LawrenceG> they must be headed towards lisp.... a few more brackets required
[02:05:20] <SWPadnos> ewwww
[02:05:38] <SWPadnos> (though I must confess I've never used lisp)
[02:08:03] <fenn> i hope they fixed the " ".join(['foo', 'bar']) vs 'foo bar'.split(" ")
[02:09:00] <fenn> but i doubt it
[02:12:08] <SWPadnos> luckily, I'm not familiar with that problem )
[02:12:11] <SWPadnos> :)
[02:14:19] <fenn> it's just one of those things where you always get it backwards
[02:17:26] <SWPadnos> oh. those look like opposite operations to me
[02:27:58] <jepler> I wish they'd have "
[02:28:15] <jepler> fixed" division so that a module could elect to have 1/3 give a rational type instead of a float
[02:28:35] <fenn> yes floats are evil
[02:28:38] <cradek> ah, ratios would be cool
[02:28:53] <cradek> does any recent system do that?
[02:29:02] <cradek> language, not system
[02:29:02] <jepler> (what I mean is, whatever the module author wants -- others might want decimal, and obviously some are happy with floats)
[02:29:32] <jepler> beats me
[02:29:43] <fenn> besides maxima?
[02:29:53] <fenn> (computer algebra system)
[02:30:02] <cradek> some lisps do
[02:30:47] <cradek> [1]> (setq r (/ 1 3))
[02:30:47] <cradek> 1/3
[02:30:47] <cradek> [2]> (* r 2)
[02:30:47] <cradek> 2/3
[02:31:07] <fenn> what happens with (+ r 0.1)
[02:31:25] <cradek> [4]> (+ r (/ 1 10))
[02:31:25] <cradek> 13/30
[02:31:32] <cradek> [5]> (+ r 0.1)
[02:31:33] <cradek> 0.43333334
[02:32:49] <cradek> (this is GNU CLISP)
[02:33:26] <jepler> apparently ruby would let you redefine 'def /' in the context of your module to make division of two ints give a rational. http://www.ruby-doc.org/docs/ProgrammingRuby/html/tut_expressions.html
[02:33:27] <SWPadnos> notation (like I) polish reverse
[02:33:47] <SWPadnos> huh, RPN backwards is NPR :)
[02:36:03] <fenn> SWPadnos: (are (doing (you (it 'wrong)))
[02:36:11] <SWPadnos> of course
[02:36:24] <SWPadnos> but not all RPN languages/systems need all the parens
[02:36:58] <jepler> lisp isn't rpn at all
[02:37:12] <SWPadnos> no, it's postfix
[02:37:15] <SWPadnos> or is it infix?
[02:37:28] <fenn> neither?
[02:37:29] <SWPadnos> it's been too long since I messed with that stuff
[02:37:32] <SWPadnos> outfix
[02:37:38] <SWPadnos> needsfixed
[02:37:55] <SWPadnos> well, not that I miss it, but too long for me to remember it much :)
[02:38:13] <fenn> Prefix notation (also known as "Polish notation"): + X Y
[02:38:39] <SWPadnos> RPM being X Y + or postfix
[02:38:42] <SWPadnos> RPN that is
[02:41:01] <jepler> oh boy, this is exciting: "Clojure - a Lisp dialect which runs on the Java Virtual Machine and has access to Java libraries."
[02:41:12] <jepler> I also heard someone tell me that there's a language called 'L#' for .net
[02:41:21] <jepler> (though I think the name should have been '(#))
[02:41:27] <SWPadnos> heh
[02:41:33] <SWPadnos> I clojed the door on that one
[02:42:03] <SWPadnos> bot.net, the easy-to-use virus-making library for .net
[02:46:29] <jepler> "clojed source softwre"
[02:46:30] <jepler> ?
[02:51:45] <SWPadnos> jepler: here's the scope trace of many many step/dir transitions: http://www.cncgear.com/images/print_03.png
[02:52:02] <SWPadnos> the fuzz is from noise over the course of 20 minutes or more
[02:52:30] <SWPadnos> the yellow and green traces are the last ones captured, and the gray is the record of all previous captured traces
[02:53:15] <SWPadnos> so it looks basically like the one you had from halscope, only with more fuzz :)
[02:56:09] <SWPadnos> oh, so that screenshot includes many keyboard jogs in both directions, plus at least a few hundred, but I think a couple ot thousand, direction reversals from a G-code program run
[02:56:52] <SWPadnos> on another note, I have verified to my satisfaction that you can move stepconf files around and it will still "do the right thing"
[02:57:23] <SWPadnos> I did modify the strings pointing to the md5sums (to be ~/emc2/... instead of /home/steve/emc2/...)
[02:57:56] <SWPadnos> I created a new user, copied the file to ~newuser, and as the new user, opened the .stepconf file
[02:58:30] <SWPadnos> and it wrote the config to the correct place under the newuser dir
[02:58:32] <SWPadnos> so yay! :)
[02:58:39] <cradek> yay
[03:00:01] <SWPadnos> I think that the timing settings provided to stepconf aren't as critical as I had believed
[03:00:18] <SWPadnos> with the exception of latency - that one is very important
[03:21:45] <jepler> do you think the step pulse timing represents a bug?
[03:22:09] <SWPadnos> regarding the dir change plot?
[03:22:18] <SWPadnos> or in stepconf?
[03:22:36] <jepler> the scope trace
[03:22:53] <SWPadnos> well, I think there may be a bug there, but I don't think it's a critical one
[03:23:13] <SWPadnos> it's a reduction in speed/efficiency, not incorrect operation
[03:23:43] <SWPadnos> but it bothers me that I can't tell you why stepgen would do that (assuming it is the fault of stepgen)
[03:25:35] <SWPadnos> I wonder if it's a servo cycle thing
[03:25:49] <jepler> the velocity command only changes every servo cycle, obviously
[03:25:57] <SWPadnos> sure
[03:26:11] <jepler> at least some of it (maybe a substantial fraction?) is in accelerating from 0 to the first step
[03:26:23] <SWPadnos> it's hard to know whether a single step should be output at the beginning, the end, or somewhere in the middle of the cycle
[03:26:53] <SWPadnos> for a jog of 0.0001 I'd expect the movement to be done in one servo cycle
[03:27:27] <SWPadnos> but stepgen also controls the direction output, so it must already know there's motion in "the other direction" before the output gets changed
[03:27:47] <jepler> talking of stepconf, you verified that you still get the "changed" warning when the file's there but user-modified too?
[03:27:58] <SWPadnos> heh - no :)
[03:28:07] <SWPadnos> there's no warning when the files are missing
[03:28:38] <jepler> there are 3 cases: missing (don't warn), present but matching (don't warn), present but different (warn)
[03:28:45] <jepler> I'm trying to figure out which cases you checked
[03:28:45] <SWPadnos> yep
[03:28:51] <SWPadnos> 1 and 3, but not 3
[03:28:54] <SWPadnos> err
[03:28:56] <SWPadnos> 1 and 2, but not 3
[03:29:12] <jepler> 3 is important, that's the point of the check
[03:29:19] <jepler> (he said, knowing he was saying the obvious)
[03:29:27] <SWPadnos> heh
[03:29:39] <SWPadnos> ok, booting that machine. I'll check it in a minute
[03:29:53] <jepler> bbl, time to do the dishes
[03:29:57] <jepler> then probably to quit for the night
[03:30:14] <SWPadnos> heh
[03:33:58] <SWPadnos> yes, case 3 works as expected
[06:03:59] <seb_kuzminsky> 'allo
[06:04:09] <seb_kuzminsky> i'm excited because i actually used my mill tonight :-)
[06:04:20] <Lerman_______> Lerman_______ is now known as Lerman
[06:04:36] <seb_kuzminsky> i made a little mounting plate to put my nice tail light on my bike
[06:04:44] <seb_kuzminsky> all in manual mode unfortunately :-(
[06:06:46] <Lerman_______> Lerman_______ is now known as Lerman
[14:20:42] <Guest113> Guest113 is now known as skunkworks_
[16:24:53] <tissf> hi
[16:26:14] <BigJohnT> hi tissf
[16:28:18] <tissf> Are there settings to adjust with the new "max feedrate" slider under the trunk version?
[16:28:45] <BigJohnT> in the ini?
[16:29:53] <tissf> yes or not ? I try to explain the problem
[16:30:09] <BigJohnT> ok, but I have to work from memory here :)
[16:30:37] <tissf> no problem John :)
[16:31:56] <tissf> after a fresh cvs from trunk on my PC workshop
[16:32:17] <tissf> I notice the new slider
[16:33:07] <tissf> I made the homming without problem
[16:34:49] <tissf> I lauch the program and... the feedrate's 645 times to low
[16:35:13] <BigJohnT> was the slider set on 100?
[16:35:21] <tissf> the max slider value is 1140mm/mn
[16:35:47] <tissf> Vel: (in axis) is 1.767 mm/mn !
[16:36:24] <BigJohnT> so your normal max feedrate is?
[16:37:12] <tissf> about 1200 mm/mn
[16:37:48] <tissf> John with the same .ini the 2.2.7 version work normaly
[16:38:21] <tissf> works
[16:38:55] <tissf> normally !
[16:39:07] <BigJohnT> hmmm, I don't see any thing to set that from here... perhaps someone who knows more about it will chime in
[16:39:21] <tissf> I'm sorry for my poor english
[16:39:33] <BigJohnT> * BigJohnT says HELP!
[16:39:52] <BigJohnT> your english is much better than my french...
[16:40:15] <tissf> that excuses nothing
[16:40:47] <BigJohnT> :)
[16:41:48] <tissf> nobody else has my speed problem?
[16:42:32] <BigJohnT> I've not upgraded mine in a few weeks so I don't know if it gives me a problem
[16:43:13] <BigJohnT> someone must have the same problem as you... now to find them...
[16:43:28] <tissf> I had not upgraded since the workshop about a month
[16:43:41] <tissf> to cold!
[16:43:51] <BigJohnT> yes, getting cold here too
[16:45:34] <tissf> 9┬░Celsius today
[16:45:55] <BigJohnT> * BigJohnT looks that up
[16:46:11] <jepler> tissf: it's probably chris's bug
[16:46:32] <cradek> tissf: on the contrary, I bet jepler did it
[16:46:38] <jepler> heh
[16:46:50] <BigJohnT> yea, jeff and chris to the rescue
[16:46:50] <jepler> the actual bug is that the french failed to get the americans to adopt metri
[16:46:51] <jepler> c
[16:47:02] <BigJohnT> 48F
[16:47:30] <jepler> BigJohnT: by "rescue" do you mean "point fingers"?
[16:48:14] <BigJohnT> jepler: what ever it takes to pass the buck when I don't have a clue and have run out of guesses
[16:49:38] <tissf> jepler: cradel: thank you I have to wait you agree between you :)
[16:50:13] <jepler> cradek: hm, is it a bug that the speed limit doesn't apply to jogs?
[16:50:38] <tissf> cradek sorry
[16:50:40] <cradek> jepler: I don't have a strong opinion on that one
[16:51:07] <tissf> jog works fine, homming to
[16:51:07] <cradek> yuck, I see maxvel is off on metric metric configs by a factor of 25.4^2
[16:51:21] <tissf> G0 and G1 bug
[16:55:17] <jepler> cradek: I've about got it fixed
[16:55:28] <cradek> oh good, because I'm still flailing
[16:56:25] <CIA-42> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix metric max velocity problem reported by tissf
[16:56:31] <cradek> thanks!
[16:56:38] <jepler> tissf: thanks for the report
[16:57:40] <tissf> oh fullspeed debugging, I am very impressed
[16:58:06] <tissf> thank you jepler: or cradek: ??? :)
[16:59:58] <cradek> haha
[17:04:04] <cradek> I confirm that fixed it - thanks jeff
[17:27:53] <seb_kuzminsky> good morning
[17:28:06] <cradek> hi seb
[19:31:05] <cradek> wow, I figured the arc problem was long gone.
[19:32:47] <cradek> SWPadnos: it's a real problem, somewhere I have a screenshot someone else sent a long time ago
[19:40:42] <cradek> aha - found it! http://timeguy.com/cradek-files/emc/G2_error.png
[19:47:58] <SWPadnos> ok, interesting
[19:48:12] <cradek> I hope he can get some information from it.
[19:48:25] <SWPadnos> yep. the version may be critical
[19:48:36] <cradek> not just the version, the exact state is critical
[19:48:44] <cradek> it's an extremely unusual problem
[19:48:57] <SWPadnos> yep
[19:49:03] <cradek> he has seen it twice, ever, with daily use of the machine for a few years?
[19:49:35] <SWPadnos> sounds that way
[19:49:50] <SWPadnos> unless there were tiny arcs where the problem wasn't noticed
[19:50:53] <cradek> haha
[19:50:59] <cradek> "how do I rewind my paper tape easily?"
[19:51:12] <SWPadnos> attach a hand crank and ... :)
[19:51:13] <cradek> "put a spool on an electric drafting eraser"
[19:51:32] <SWPadnos> they made electric erasers?
[19:51:38] <cradek> dude
[19:51:38] <jepler> oh sure
[19:51:43] <cradek> of course they did
[19:51:47] <cradek> but, good luck finding one
[19:51:55] <SWPadnos> I guess I never made that many drafting mistakes :)
[19:52:08] <SWPadnos> my Stadtler mars plastic eraser was enough
[19:52:09] <cradek> I just think it's funny advice for a funny problem
[19:52:13] <SWPadnos> ineed
[19:52:15] <SWPadnos> +d
[19:52:34] <jepler> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=180310297926
[19:52:59] <SWPadnos> cool
[19:53:03] <jepler> though I'd have to recommend the corded one: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300277518739
[19:53:11] <SWPadnos> and they have 5 of them at that low low price
[19:53:13] <jepler> The eraser works at excellent speed (not sure how much).
[19:53:39] <SWPadnos> heh - "not recommended for fine art work"
[19:54:30] <fenn> CNC eraser anyone?
[19:55:00] <SWPadnos> so you can erase the evidence when your arcs end up being lines?
[19:55:13] <jepler> I bought 3 brushless motors, now can someone tell me how to connect them for cnc? http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=130269028360
[19:56:06] <jepler> "Bonded needymium(sic) magnet" ?
[19:56:09] <SWPadnos> http://www.chorusmotors.com/demo/index.demo.html
[19:56:14] <SWPadnos> (sorry, it's flash)
[19:59:39] <jepler> interesting
[20:00:28] <SWPadnos> yeah - it seems like a cool technology (and it has good descriptions of deta and wye connected 3-phase motors too)
[20:01:13] <SWPadnos> I'm glad they're finally working on putting motors in landing gear too - that's something I thought of a few years ago (but couldn't borrow a jumbo jet to test with :) )
[20:19:42] <micges> how to add kinematics to emc?
[20:39:35] <micges> ok I find it in makefile
[20:49:03] <BigJohnT> what does the and2 function and2.all do?
[20:57:07] <SWPadnos> if you have more than one and loaded (count>1 or multiple names), it run all of them
[21:01:41] <BigJohnT> like setp and2.all.in0 1
[21:02:20] <SWPadnos> no, like addf and2.0 fast / addf and2.1 fast / addf and2.3 fast ...
[21:03:02] <SWPadnos> it's like using parport.write-all instead of adding parport.0.write and parport.1.write and parport.2.write (...) to a thread
[21:03:29] <SWPadnos> the advantage is that it's one function, and all the "and2" blocks are evaluated at once
[21:04:12] <SWPadnos> the disadvantage is that you can't decide on the sequencing of their evaluation
[21:05:07] <SWPadnos> it's also not ideal if you want to have some of them get evaluated in the servo thread (for user or CL-related stuff maybe) vs. in the base thread (for combining multiple limit switches, for example)
[21:10:32] <BigJohnT> ok, thanks that makes sense now
[21:15:43] <micges> applying perpedicularity correction to kinematics work perfectly
[21:15:54] <SWPadnos> cool
[21:16:06] <SWPadnos> should be pretty simple math too :)
[21:16:59] <micges> math is trivial but real behavior of machine will be tested on table mill on monday
[21:18:13] <alex_joni> micges: it would be good if you would consider submitting that for inclusion
[21:20:20] <micges> sure
[21:21:09] <micges> after testing on monday
[21:22:02] <micges> one thing: what can I do when I want let say change scewness from 0.1 to 0.05
[21:22:17] <micges> joint following error will hapen
[21:22:41] <alex_joni> maybe you can do that with machine switched off
[21:22:44] <alex_joni> F2
[21:23:31] <micges> ok
[21:24:37] <SWPadnos> or move to 0,0,0 in joint mode
[21:24:47] <alex_joni> that might be a bit harder :)
[21:25:11] <SWPadnos> heh
[21:25:19] <SWPadnos> it sounded good though :)
[21:25:22] <SWPadnos> brb
[21:26:08] <alex_joni> might be something :)
[21:26:57] <alex_joni> bbl
[22:09:14] <micges> good night
[22:21:51] <alex_joni> cradek: maybe dahlihvar.txt is ok?
[22:22:01] <SWPadnos> it looks fine to me
[22:22:06] <SWPadnos> just some offests there
[22:22:44] <alex_joni> what would you have expected?
[22:22:52] <SWPadnos> nothing in particular
[22:22:57] <alex_joni> nan's and inf's :D (debugging doesn't get that easy :P)
[22:23:01] <SWPadnos> heh
[22:23:13] <SWPadnos> # the following line may contain an error
[22:23:16] <SWPadnos> :)
[23:02:43] <CIA-42> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gui/axis.lyx: add special comments section
[23:11:20] <alex_joni> good night all
[23:18:32] <cradek> darn, wish he was here.
[23:18:37] <cradek> I would so much like to nail this bug
[23:38:50] <BigJohnT> I have a nail gun you can borrow cradek
[23:40:47] <CIA-42> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/basic_hal.lyx: add more info
[23:41:56] <CIA-42> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/hal/basic_hal.lyx: add more info