Guest User

Untitled

a guest
Aug 21st, 2018
110
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 7.33 KB | None | 0 0
  1. <enleth> so I had this idea a couple years ago to mount the toolhead on a non-contact plotter-like gantry machine (such as a plasma cutter) on a small, limited range X-Y stage, in addition to the regular X (gantry) and Y (linear stage on the gantry) axes
  2. <enleth> or perhaps just an additional linear stage parallel to X
  3. <enleth> the idea here being to split the toolpath in frequency domain and use the small stage for the high frequency part
  4. <enleth> never tried it, mostly for the lack of software support in pretty much any machine control out there
  5. <enleth> and now I'm wondering how much work would it take to get LinuxCNC to understand and use such a configuration
  6. <enleth> obviously the whole point is being able to run detailed toolpaths fast on a machine that would otherwise lack the rigidity and/or power to do that
  7. <enleth> any thoughts?
  8. <_unreal_> What are you asking
  9. <Jymmm> CaptHindsight: It is I, no promises on the latter ;)
  10. * Patang ([email protected]) has joined #linuxcnc
  11. <enleth> _unreal_: high-level comments, especially from people familliar with LinuxCNC's motion planner and HAL. I assumed this is obvious when asking for "any thoughts" after describing a high-level idea dependent on the internals of the motion planner and HAL.
  12. * phiscribe_ ([email protected]) has joined #linuxcnc
  13. <_unreal_> enleth, AGAIN.... WHAT are you asking I've read it multi times now...... your question is not really narrow
  14. * phiscribe has quit (Ping timeout: 272 seconds)
  15. <_unreal_> I assume you mean you want to setup a TOOL rotation system so you can swap tools. THAT IS NOT an XY setup. but an A rotation style axis generally. unless you want to use some kind of a PICK and PLACE system to pick up a TOOL from a XY matrix style tool holder....
  16. <_unreal_> SOOO... narrow the scope of WHAT your are asking
  17. <enleth> _unreal_: jesus, not everything said here has to *be* a narrow question. if you really insist, the question is: "what do you guys think, is that even remotely doable within the LinuxCNC codebase as it is now, or am I hosed"?
  18. <enleth> _unreal_: no, you completely misunderstood it
  19. <_unreal_> I may have no argument
  20. <_unreal_> brb OMG 9:40 putting my 9yo to bed
  21. <_unreal_> one sec
  22. <_unreal_> I'm back
  23. <_unreal_> kids to bed
  24. <enleth> _unreal_: this is about doing fast moves on a heavy or wobbly machine, by being able to move the tool itself relative to the toolhead, separately from the machine's "regular" axes, with a very limited range of motion but at a potentially much higher speed than the "regular" axes can do
  25. <_unreal_> What kind of a mchine
  26. <_unreal_> machine
  27. <enleth> gantry plotter comes to mind, but this really isn't that important
  28. <_unreal_> stationary or moving gantry
  29. <enleth> that's utterly, completely irrelevant
  30. <enleth> I'm not asking about a particular machine
  31. <_unreal_> it is important because every moving part ADD'S the error range of play
  32. <enleth> I'm getting the impression that we're not quite on the same page here
  33. <_unreal_> fine ok.. hypothetical a plotter with a DUAL X plane
  34. <veek> anyone familiar with the Aluminium Classification system?
  35. <_unreal_> I'm having a hard time picturing an XY/XY table
  36. <enleth> what I'm asking about is a motion control idea/model/concept, how the machine looks makes absolutely zero difference
  37. <_unreal_> sorry XY/XY flimsy fast plotter
  38. >veek< AM i WORNG?
  39. <_unreal_> veek, are you asking about its grade of duriblity and MOMENT?
  40. <_unreal_> sorry flex and moment
  41. <veek> nope the whole 1xxx implies 99%
  42. <veek> what does the 2nd digit etc imply
  43. <_unreal_> duriblity lol I've had to much wine and I'm on my first cup :)
  44. <_unreal_> OHOHHHHHHHHH.... ya that I cant answer
  45. <enleth> _unreal_: it's probably better if you imagine a VERY heavy machine that would need to slow down a lot to trace a complex toolpath with very small details, like a high amplitude sine wave set diagonally to the axes
  46. <_unreal_> I'd have to google
  47. <_unreal_> enleth, ok... microstepping.... on a machine with high backlash
  48. <enleth> _unreal_: with a dual X/Y stage, one of them very small and right at the toolhead, the machine as a whole could zoom over the toolpath in a straight line, with only the tool itself doing the sine wave
  49. <enleth> jesus, no
  50. <veek> _unreal_, https://www.esabna.com/us/en/education/blog/understanding-the-aluminum-alloy-designation-system.cfm
  51. <enleth> what microstepping
  52. <enleth> what backlash
  53. <enleth> completely missing the point
  54. <_unreal_> veek MSG
  55. <veek> good explanation my book sucks
  56. <_unreal_> veek, check msg
  57. <enleth> _unreal_: I'm sorry, I tried, but I don't think I'm getting anything across now. and my whole question is a long way up in the backlog and no one else is going to read it and comment on it
  58. <enleth> what a waste of time
  59. <_unreal_> I'm trying. I'm a marine engineer
  60. <_unreal_> but the question that I'm reading I've answered as best I can..... as I said its best to narrow your question and build up.......
  61. <_unreal_> your asking a question that is NOT making sense.
  62. * sector_0 ([email protected]) has joined #linuxcnc
  63. <enleth> the question was about the LinuxCNC software internals and control theory, NOTHING to do with the mechanical implementation
  64. <_unreal_> linuxcnc is very flexable and can be scripted.......
  65. <enleth> not at that level, I think
  66. <enleth> and I know it is, I'm using it
  67. <_unreal_> if you feel that the lincnc is not capable then you need to take it to a hardware level abstraction layer and use a micro controller to preform automated functions
  68. <enleth> what
  69. <enleth> again, what you said has nothing to do with anything
  70. * CaptHindsight has quit (Quit: Leaving)
  71. <_unreal_> http://linuxcnc.org/docs/html/man/man1/iocontrol.1.html
  72. <enleth> yes, I know where the manual is, this is irrelevant
  73. <_unreal_> At this point I'm sorry I give up... I dont know what your desired answer is between what your writing and thinking sorry.
  74. * dirtyroshi ([email protected]) has joined #linuxcnc
  75. * Patang has quit (Ping timeout: 244 seconds)
  76. * phiscribe ([email protected]) has joined #linuxcnc
  77. <enleth> _unreal_: you're a marine engineer, right? so try to imagine someone asked you to draw a sine wave - 2m in amplitude, 4m period - with a stick on water, while you're standing at the stern of a 80000 DWT tanker. you're holding the stick in your hands, it's long enough to reach the water. would you rather try to get the whole tanker to strafe left and right by 2m as it's moving forward, or just let the
  78. <enleth> ship move in a straight line and wave the comparatively lightweight stick with your hands so that the end in the water traces a sine wave?
  79. * phiscribe_ has quit (Ping timeout: 272 seconds)
  80. * dirtyroshi has quit (K-Lined)
  81. <enleth> the end result would be exactly the same if you held the stick straight and somehow got the ship to move sideways at a 2m amplitude, 2m period, but surely you can see a problem in trying to do that with a tanker
  82. <enleth> even if you tried using side thrusters, the mass of the ship would just completely dampen that sideways movement, or you'd have to move the whole thing *extremely* slowly
  83. <enleth> eh, now no one is going to comment anything on what I originally wrote because no one's going to read it anymore
  84. <enleth> _unreal_: and, yes, I am aware that a line traced in the water with a stick while standing on a big ship's stern is going to disappear right into the ship's wake, let's just assume it doesn't
Advertisement
Add Comment
Please, Sign In to add comment