#linuxcnc-devel Logs

Jul 11 2018

08:28 AM JT-Shop: pcw_home: what is the link to the 7i96.zip file?
08:35 AM skunkworks: JT-Shop: aww - you get to play with all the fun stuff :)
08:40 AM pcw_home: http://www.mesanet.com/software/parallel/7i96.zip
08:58 AM JT-Shop: thanks, now to figure out what they are :)
09:04 AM * JT-Shop heads to work
04:24 PM hazzy-lab: Can joint units ever differ from machine units??
04:28 PM hazzy-lab: More specifically, [JOINT_n] UNITS can be used to override [TRAJ] UNITS, but how does that work?
04:32 PM hazzy-lab: I am trying to clarify some of the docs ...
04:42 PM hazzy-lab: Ok, it appears that [JOINT_n] UNITS is ignored
04:42 PM hazzy-lab: (at least stat.joint[n]['units'] returns machine units even if I change the units setting for JOINT_n)
04:45 PM hazzy-lab: I don't understand how the kinematics modules work, but in theory could somebody implement a kins module that could use different units for each joint? (as crazy as that would be :D)
04:50 PM andypugh: Not seen this before: “Note: Using LXRT Realtime”
04:50 PM andypugh: The user is complaining of poor step smoothness, and seems to have an RTAI kernel
05:08 PM cradek: but maybe not the rtai linuxcnc?
05:09 PM seb_kuzminsky: i think lxrt is the userspace realtime that rtai offers
05:09 PM seb_kuzminsky: could it be they're running a uspace build on an rtai kernel?
05:10 PM cradek: src/rtapi/uspace_rtai.cc: rtapi_print_msg(RTAPI_MSG_ERR, "Note: Using LXRT realtime\n");
05:11 PM seb_kuzminsky: grep > guessing
05:13 PM andypugh: Yes, I think that is it. As it isn’t working very well I have told them to run configure again with realtime set to RTAI. I hope that is what it takes.
05:13 PM seb_kuzminsky: that sounds like good advise
05:14 PM andypugh: Has anyone (Purposely) experimented with LXRT?
05:16 PM seb_kuzminsky: i think only jepler
05:33 PM cradek: http://timeguy.com/cradek/01531345785
05:34 PM hazzy-lab: Any input??
05:34 PM hazzy-lab: https://github.com/LinuxCNC/linuxcnc/pull/462/commits/69426707799e0cb620094e2c529769e9bfd2b307
05:39 PM andypugh: Interesting wrinkle
05:46 PM hazzy-lab: andypugh: It is!
05:46 PM hazzy-lab: The docs mention that TRAJ units can be overridden by JOINT units, but in my limited testing the JOINT units seem to be ignored
05:47 PM hazzy-lab: It would be kinda neat to be able to have separate joint units!
05:51 PM hazzy-lab: You could make the joint units match the screw units, which would alleviate the need for rather complicated unit conversions :D
05:54 PM andypugh: Is that what would happen?
05:54 PM andypugh: I am not convinced.
05:54 PM andypugh: In fact I am not particluraly sure what it would mean
05:55 PM andypugh: I guess it can only (meaningfully) be a part of the scale calculation.
05:56 PM hazzy-lab: I _think_ all units are converted to mm, so it would just change the conversion factor for the values in that joint?
05:57 PM hazzy-lab: Scale, but also things like limits would be effected
05:58 PM hazzy-lab: (or is that affected? lol)
06:01 PM hazzy-lab: I can't think of any situation were it would be *necessary* to have the joint units different from the traj units ...
06:25 PM skunkworks: cradek: cool!
06:25 PM skunkworks: mine is still humming
06:48 PM skunkworks: cradek: how is the coil winder working?
06:55 PM skunkworks: http://electronicsam.com/images/greenmachine/20180710_190604.jpg
06:55 PM seb_kuzminsky: andypugh: does that commit apply in 2.7? i bet 2.7 has the same bug
06:55 PM seb_kuzminsky: skunkworks: dang that's a big cutter!
06:57 PM andypugh: seb_kuzminsky: I am not sure it really counts as a bug.
06:57 PM skunkworks: 6"
07:55 PM Tom_L: quite the fixture too :)
08:41 PM cradek: skunkworks: it works great. it's not super fast, but I also haven't tried to speed it up. about 30 minutes/coil.
08:42 PM cradek: skunkworks: I'm super happy to be able to do it, but each one still represents a whole lot of work
08:45 PM cradek: that is a large flycutter. I see a dovetail at the bottom. what are you doing?