#linuxcnc-devel Logs

Jun 08 2018

12:41 PM pcw_mesa: Is there some good reason for motion.requested-velocity to be any value except the requested feed rate or the rapid rate?
12:43 PM pcw_mesa: It seems to have random odd values
12:44 PM seb_kuzminsky: i dont know if that pin is supposed to have the vel that the user requested of motion (in which case i agree with you, modulo feed overrides), or the vel that motion requested of the machine (in which case it could be any value up to those limits, as dictated by g64 etc)
12:49 PM pcw_mesa: The man page just saye requested velocity (before any feed rate modulation)
12:49 PM pcw_mesa: says
12:50 PM seb_kuzminsky: the code shows that pin has the velocity of the current TP segment, so i'd expect it to be below the commanded velocity by whatever amount the trajectory planner needs to keep the tool on the path
12:50 PM seb_kuzminsky: the motion manpage is... pretty terse
12:51 PM seb_kuzminsky: does my reading of the code match your observations of the unexpected pin values?
12:51 PM pcw_mesa: I think so
12:52 PM seb_kuzminsky: i've got to run, i'll read back tonight
12:52 PM pcw_mesa: There is a use case for the gcode requested velocity though
12:52 PM seb_kuzminsky: ah, ok, maybe we can add that
12:52 PM pcw_mesa: (plasma THCs so you can set the corner lock velocity)
12:52 PM seb_kuzminsky: can you open a github issue with details and i'll look into it?
12:52 PM seb_kuzminsky: for the feature request, i mean
12:52 PM seb_kuzminsky: bbl
12:52 PM pcw_mesa: (as a percent of the current feed rate)
12:53 PM pcw_mesa: OK will do
02:51 PM hazzy-lab: state tags! is anybody working on it? It would be real nice to have merged
02:51 PM hazzy-lab: I was going to test it a while back but could not get it to compile
02:51 PM hazzy-lab: but I did not try real hard :)
03:03 PM skunkworks: for now i think you could send the feedrate to hal using something like m68 e0 q'feed'
03:08 PM skunkworks: could be an input filter
03:13 PM * skunkworks acts like he knows what he is talking about
03:29 PM pcw_mesa: Yeah not sure what the tradeoffs are of using M67.68 (other than making the gcode LinuxCNC specific, but that may be the case anyway)
03:31 PM * pcw_mesa didnt read the filter commnet
03:33 PM pcw_mesa: I think the state tags would be nice here but realistically for plasma a hacked in motion pin is probably adequate
05:48 PM hazzy1 is now known as hazzy
10:55 PM jepler: hm I have a kunena update to do that I've been putting off. But it's 5.0.14 -> 5.1.0 and that trailing dot zero is always worrysome
10:55 PM jepler: I plan to do it in the morning (US CST)
10:55 PM jepler: so I'll be awake when something goes wrong