Back
[00:24:17] <linuxcnc-build> build #3483 of 0000.checkin is complete: Failure [4failed] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3483 blamelist: andypugh <andy@bodgesoc.org>
[11:13:56] <REEEN> Hello Guys
[11:14:05] <REEEN> Anybody online ?
[11:14:19] <seb_kuzminsky> hi REEEN
[11:14:30] <REEEN> hello seb
[11:14:50] <REEEN> I have a question, maybe you can tell me the answer
[11:14:54] <seb_kuzminsky> maybe
[11:15:23] <REEEN> I made a config for a 5 axis machine, one axis is a rotary so I set wrapped rotary to 1
[11:15:52] <REEEN> the wiki says that it bahve that way :
[11:16:23] <REEEN> g1 C90 : c axis will move ccw direction
[11:16:41] <REEEN> g1 c-90 : c axis will move cw direction
[11:17:05] <REEEN> but somehow it will always turn ccw ??
[11:17:31] <REEEN> and when I command c90 and then c-90 it doesn't do anything
[11:17:53] <REEEN> homing works fine :D
[11:17:58] <cradek> if you're at c90, another c-90 should not cause motion
[11:18:11] <cradek> I don't understand what you mean by 'somehow always ccw'
[11:18:54] <seb_kuzminsky> doesnt C-90 mean the same as C[360-90], so there should be motion?
[11:19:11] <REEEN> yes look here :
[11:19:25] <seb_kuzminsky> g0 c90 ; move in positive direction to C=90
[11:19:30] <cradek> g90c90 and g90c-90 are the same destination
[11:19:39] <REEEN> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?WrappedRotaryAxes
[11:19:40] <seb_kuzminsky> g0 c-90 ; move in negative direction to C=270
[11:19:45] <cradek> no
[11:19:55] <cradek> move in negative direction to C=90
[11:19:59] <cradek> but you are already there
[11:20:33] <REEEN> 3. Effect of rotary motion in wrapped rotary motion mode When a rotary move is commanded, the sign ("+" or "-") of the axis word recorded. Then, any offsets in effect are added to the absolute value of this angle to give the commanded angle 'theta'. The motion is to an angle equivalent to theta If the sign of the axis-word is "+", then the motion is counterclockwise If the sign of the axis word is "-", then the motion is clockwis
[11:20:57] <cradek> I don't know if this draft spec is exactly right
[11:21:06] <cradek> I don't understand what you are doing that you think gives a wrong result
[11:21:22] <cradek> can you please give example gcode, what you expect it to do, and what it does instead?
[11:21:30] <REEEN> yes wait
[11:22:34] <cradek> (I wonder if any UI other than AXIS shows the right DRO values...)
[11:22:48] <cradek> brb
[11:23:02] <REEEN> the draft says that any commanded motion with a positive sign will make the axis move in ccw direction and any commanded motion with a negativ sign will result in a cw direction
[11:23:40] <REEEN> so g1 c90 means to me : move 90 degree ccw
[11:24:00] <REEEN> and g1 c-90 means move -90 degrees
[11:24:16] <cradek> no
[11:24:26] <archivist> I have read the wrapped docs and dont use it because it seems wrong
[11:24:28] <cradek> to destination = 90
[11:24:43] <REEEN> so the docs are totally wrong
[11:24:56] <cradek> I think you are describing g91 behavior
[11:25:06] <REEEN> can somebody please correct that ?
[11:25:19] <REEEN> in the docs they never say anything about g91
[11:26:00] <cradek> is this even described in the real (non wiki)l docs?
[11:26:15] <jepler> > When a rotary move is commanded, the sign ("+" or "-") of the axis word recorded. Then, any offsets in effect are added to the absolute value of this angle to give the commanded angle 'theta'.
[11:26:45] <jepler> the "absolute value" of -90 is 90, which is why C-90 moves to the angle 90, not 270.
[11:27:01] <cradek> it's too bad that documentation describes implementation, not behavior
[11:27:12] <cradek> no wonder it's confusing
[11:27:26] <jepler> I guess an implementation of this actually went in linuxcnc?
[11:27:35] <cradek> yes years ago
[11:28:42] <REEEN> so how can I move my axis cw and how ccw with degrees ?
[11:30:03] <cradek> I don't understand 'with degrees', can you ask a different way?
[11:30:31] <REEEN> I want my axis to move 90 degree ccw what do I do ?
[11:30:37] <cradek> (I wish the online docs had a search)
[11:30:43] <cradek> g91 g0 c90
[11:31:06] <REEEN> okay so I have to use g91 ?
[11:31:24] <SEL> hello, i tried this command G1 A90 moves 90 degrees with relative coordinate command before it
[11:31:25] <cradek> yes if you want a move with destination relative to where you currently are, that's g91
[11:32:48] <REEEN> the problem is
[11:33:06] <REEEN> my cam will move my x y z and b axis in g90 commands
[11:33:23] <REEEN> and it will move my c axis with g91 :D
[11:34:33] <REEEN> SO I can't say G1 x50 c90 because my cm means : G90 G1 X50 G91 G1 C90 in one line
[11:35:14] <cradek> no, you can't have g90 and g91 in the same line
[11:35:42] <cradek> sounds like you either need to fix your cam postprocessor or choose a linuxcnc mode that's compatible with it
[11:36:25] <jepler> to the best of my knowledge, LinuxCNC doesn't have any way to say: Move X to absolute coordinate 1, and move C by relative coordinate 2
[11:36:32] <jepler> .. as a single coordinated move
[11:36:37] <cradek> nope
[11:36:42] <cradek> you just can't do that in gcode
[11:36:54] <REEEN> yes that's the problem
[11:37:05] <cradek> are you sure that's what your cam means? that seems very unlikely to me
[11:37:26] <cradek> very unlikely because I think no gcode dialect works that way
[11:37:41] <REEEN> It doesn't say g90 and g91 in one line but it wants to move xyz absolut and c incremental
[11:38:01] <REEEN> I will try that out
[11:38:07] <REEEN> One last thing :
[11:39:27] <REEEN> I did this as a test G90 G1 C90 => ccw move 90 deg and then G90 G1 C0 and then it also moves ccw to 0 degress in a 3/4 turn ?
[11:39:39] <jepler> yep
[11:39:58] <jepler> C0 is the same as C+0, so the rotation will be in the positive direction
[11:40:00] <cradek> yes that is the right behavior
[11:40:15] <jepler> and the destination angle is zero degrees
[11:41:17] <REEEN> what would have happened If I would move to C-20
[11:41:41] <cradek> it would move from 90 downward (cw) to destination 20
[11:41:41] <jepler> C-20 will turn in the negative direction
[11:41:46] <jepler> and it will to turn to 20
[11:42:23] <REEEN> ah okay so 0 is positiv
[11:42:38] <cradek> if you put no + or - sign it is assumed to be positive
[11:42:55] <cradek> (all of gcode works that way)
[11:42:58] <REEEN> how can I move from c90 back to c0 in negativ direction ? c-360 ?
[11:43:04] <cradek> c-0
[11:43:14] <cradek> if in g90, g0 c-0
[11:43:24] <REEEN> okay
[11:43:41] <REEEN> I will try that out
[11:43:46] <REEEN> thank you guys !!!!
[11:43:56] <cradek> welcome
[11:44:06] <REEEN> hmm those docs are so confusing
[11:44:14] <cradek> (bet we don't have test coverage for this)
[11:44:22] <cradek> where are the docs? I didn't find them
[11:44:34] <REEEN> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?WrappedRotaryAxes
[11:44:39] <REEEN> those
[11:44:43] <cradek> no I mean the real docs
[11:44:52] <cradek> that is an old spec for development
[11:45:11] <REEEN> the only one I remember is in the ini section
[11:45:28] <cradek> yeah that's all I found too
[11:45:30] <REEEN> ini axis wrapped rotary
[11:45:33] <cradek> I bet it is undocumented
[11:45:48] <REEEN> okay :)
[11:45:53] <REEEN> thanks again !
[11:46:45] <seb_kuzminsky> maybe here would be a good place to add this documentation:
http://linuxcnc.org/docs/2.7/html/gcode/machining-center.html
[11:47:51] <cradek> yes section 2.2
[11:48:21] <cradek> um, introducing the term "wrapped linear" doesn't help that be clear
[11:48:27] <seb_kuzminsky> haha
[11:51:58] <jepler> hm, here's an interesting feature attributed to a Fadal contol. G17 Q- puts you in "cam wrapping mode", in this case making Y words command A rotations based on the Q- word which is computed from the cam diameter
[11:52:33] <cradek> that's interesting
[12:00:12] <REEEN> okay guys its working as expected
[12:00:19] <REEEN> :)
[12:01:09] <REEEN> wow another thing ... i just updated to linuxcnc 2.7 and if I type sudo comp --install wzw.comp
[12:01:23] <REEEN> it says comp command not found ??
[12:03:07] <pcw_home> its halcompile now
[12:04:51] <REEEN> okay
[12:05:24] <REEEN> thanks
[12:58:48] <maxcnc> hi is the xhc handheld crew sometimes online here ?
[12:59:02] <maxcnc> frederic ,rene, marius
[12:59:58] <maxcnc> i did a i think realy improve so it can jog on the first 3 stepssizes and button move on the next 3
[13:00:29] <maxcnc> and you can select individuell numbers
[13:01:12] <maxcnc> im on meric plasma so i use 0.1 , 0.5 1 , 250,1000,3500
[13:01:14] <seb_kuzminsky> maxcnc: you're probably best off posting it to the emc-developers mailing list
[13:01:51] <maxcnc> i also hackt into the system to discover the missing icons connection show up
[13:02:28] <maxcnc> bad they made them on the same flag as the stepsize but with plausible logic
[13:02:37] <maxcnc> so you coudt sum them up
[13:02:57] <maxcnc> to show the icons with the stepsize
[13:04:31] <maxcnc> ok to my person im the man who owns Lothars shop you may know him from the past
[13:05:10] <maxcnc> we past the 250 mashine build number
[13:06:23] <maxcnc> so bye i will get more out of the xhc and come back if i get over the c++ pointers
[13:06:33] <maxcnc> BYE
[15:51:10] <seb_kuzminsky> hi skunkworks
[15:56:51] <skunkworks> Hi seb_kuzminsky
[15:56:55] <skunkworks> how is it going?
[15:58:40] <skunkworks> I know it is probably too soon to say - but 2.7 with all the changes seems to be going well.
[15:59:02] <skunkworks> could be the rose colored glasses
[17:04:57] <seb_kuzminsky> skunkworks: i agree! it took a long time to stabilize, and there are a couple of bugs open still, but overall i'm very happy with it
[17:05:41] <seb_kuzminsky> i can't repro tom easterday's bug report, can you try your magic on it?
[17:05:46] <Tom_itx> other than changing a couple ini settings i've had no issues
[18:58:56] <skunkworks> wonder if it is a re-map issue
[18:59:07] <skunkworks> (just thinking out loud)
[21:26:07] <seb_kuzminsky> i thought he was not using remap, did i misread?
[21:48:48] <skunkworks> ah - I suck at reading comprehension.. (and spelling)
[21:56:15] <seb_kuzminsky> "I could remap, but..."