#linuxcnc-devel Logs

Sep 25 2018

#linuxcnc-devel Calendar

07:15 AM KaitoDaumoto is now known as Guest56500
02:08 PM andypugh: Would nyone care to speculate on what these NML commands do? https://github.com/LinuxCNC/linuxcnc/blob/master/src/emc/nml_intf/canon.hh#L535
02:10 PM andypugh: https://github.com/LinuxCNC/linuxcnc/blob/0f91c553a238c3f5e8a52285044761c2dcfd7de5/src/emc/task/emccanon.cc#L1886
02:10 PM andypugh: Nothing.
02:10 PM andypugh: But what are they _meant_ to do, I wonder?
02:15 PM rmu: andypugh: google found this https://ws680.nist.gov/publication/get_pdf.cfm?pub_id=820407
02:16 PM rmu: The LOCK_SPINDLE_Z
02:16 PM rmu: command locks the spindle against vertical motion.
02:16 PM andypugh: The UNLOCK_SPINDLE_Z command unlocks the spindle to permit vertical motion. The SET_SPINDLE_FORCE (force) sets the force with which the spindle is pushed in the z-direction, potentially useful in tapping operations
02:16 PM rmu: The UNLOCK_SPINDLE_Z command
02:16 PM rmu: unlocks the spindle to permit vertical motion. The SET_SPINDLE_FORCE (force) sets the
02:16 PM rmu: force with which the spindle is pushed in the z-direction, potentially useful in tapping
02:16 PM rmu: operations.
02:16 PM rmu: ;)
02:17 PM andypugh: I have no idea how one would implement them. And clearly no-one has ever wanted them
02:18 PM rmu: seems to be leftovers from some NIST research
02:20 PM cradek: I bet these are all easy to understand if you think of them in terms of matt shaver's manual bridgeport that they were retrofitting in the very early days
02:21 PM cradek: those had a quill that locked and had a handle like a drill press that's very useful for drilling and tapping
02:21 PM sync: only for drilling really
02:21 PM cradek: nobody drills by calculating a feed on those, you just lean on the handle an amount that feels right and the drill drills whatever feed it likes
02:21 PM cradek: for tapping you'd start it by leaning on it, then let go, and reverse the spindle when it gets down far enough
02:24 PM Tom_L: the good ole days
02:32 PM andypugh: Spindle lock would be easy to implement, add a HAL pin
02:33 PM andypugh: But the force control? I guess another HAL pin and leave implementation to the user
02:34 PM sync: torque * pitch, but yeah
02:34 PM cradek: I think the part where you'd need to stop doing position control and then later start doing it again is the hard part
02:35 PM andypugh: Cam I stop fixing “runtests” when my branch fails the same number as master?
02:35 PM cradek: I can imagine air being a good way to set a certain force for drilling
02:39 PM sync: cradek: only switching between them is actually hard if you don't want discontinuities
02:40 PM cradek: yeah I agree, I meant to say that's what makes it hard
02:59 PM KimK: Sometimes "single-point broaching"(?) operations are done by pulling/pushing a tool vertically while keeping the spindle from rotating, and moving tiny increments in X & Y to nibble away the desired material. Could that be what they were doing?
03:00 PM KimK: I said vertically, assuming vertical mill.
03:03 PM KimK: If it was a Bridgeport (or clone) they were working with, I suppose they could lock the quill and broach with the knee?
03:07 PM andypugh: You can buy slotters for BP
05:50 PM KimK: Indeed, here are three on eBay: https://www.ebay.com/itm/323465257546 https://www.ebay.com/itm/223118902996 http://www.ebay.com/itm/223145175638
06:10 PM andypugh: I missed the correct slotting head for my mill a while ago. Unusually when the auction went for 282 and I had bid 250 I found myself wishing I had bid 350. Especially as I have three keyways to cut.
06:20 PM seb_kuzminsky: andypugh: i think the problems in master might just be a flaky arm buildslave
06:21 PM seb_kuzminsky: all the other platforms are passing
06:21 PM andypugh: Hmm, yes
06:21 PM andypugh: In my branch I have all but M98 no 12 passing
06:21 PM andypugh: That is giving Bug code not m0 m1 m2 m30 m60 m99
06:22 PM andypugh: But master is OK
06:27 PM andypugh: It seems that I am ending up in interp::convert_stop by mistake…
06:28 PM andypugh: Ah heck, I think already fixed this once…
06:28 PM andypugh: M99 is a valid stop code
06:30 PM seb_kuzminsky: isnt m99 like an o-word now?
06:30 PM andypugh: It’s a merge conflict I messed up, and seem to have reverted
06:30 PM andypugh: (I fixed it once already)