#linuxcnc Logs

Apr 23 2019

#linuxcnc Calendar

02:37 AM Loetmichel: mornin'
04:17 AM SpeedEvil: Loetmichel: lolz.
04:17 AM SpeedEvil: (at ridiculous requirements)
04:18 AM Loetmichel: SpeedEvil: would be so bad if i had a CNC lathe... that small C0 on the other hand is a PITA to work series with.
04:18 AM Loetmichel: https://www.youtube.com/watch?v=4n0rmiiDVD4
04:18 AM Loetmichel: :-)
04:19 AM SpeedEvil: Make one cylindrical part, and then slot and press a plug.
04:21 AM Loetmichel: if i had cylindrical brass stock i would
04:22 AM Loetmichel: but i only have some brass M4 standoffs... :-)
04:25 AM rmu: Loetmichel: looks like a job for a "geschützte werkstatt", don't know how to translate that into englisch
04:34 AM XXCoder: literally, protected workshop
04:34 AM XXCoder: not sure what it should mean
04:42 AM Loetmichel: XXCoder: "behindertenwerkstatt"
04:43 AM XXCoder: interesting
04:43 AM XXCoder: gonna love germany language word combiner thing
04:44 AM Loetmichel: rmu: nope, its a job for a CNC "langdreher" lathe
04:44 AM syyl: not for 30
04:44 AM Loetmichel: that can feed 6m long round stock
04:44 AM syyl: they usualy hate you for making them setup a job for such small numbers
04:45 AM Loetmichel: syyl: correct, i would have made a few 100 of that so i have some stock if i outsoruced it to a CNC shop
04:45 AM syyl: there are only a small number of swiss-lathe shops that are setup to do small runs
04:45 AM XXCoder: sub-10 parts jobs at shop i work at is rare
04:45 AM XXCoder: and tend to be very hard jobs
04:45 AM Loetmichel: i know one... they even to single "non-standard" gears if you pay...
04:45 AM syyl: thats the parts i look for, XXCoder :D
04:46 AM Loetmichel: but boss is relu8ctant to pay his exorbitant prices
04:46 AM XXCoder: shop i work at is too expensive for easy few part job lot
04:46 AM syyl: which is absolutely ok, XXCoder
04:46 AM XXCoder: syyl: guy at work just finished out 2,000 parts job lol
04:47 AM syyl: no sense in tying up a highend machine with bumblefucking around on small runs ;)
04:47 AM XXCoder: its essenally just rouned top bar. made from larger alum thin plate (3mm or so)
04:47 AM rmu: my comment was not really serious. the obvious solution in the fuse-holder-case is to change the fuse holders
04:47 AM Loetmichel: XXCoder: usually "benhindertenwerstätten" have mostly mentally disabled workers, not so much physically disabled ones
04:47 AM XXCoder: loet interesting. no sensory loss jobs?
04:47 AM XXCoder: like blinded
04:47 AM Loetmichel: yeah, they are there, too
04:47 AM Loetmichel: but not that many
04:48 AM XXCoder: lighthouse here is all visual problems
04:48 AM Loetmichel: blind people and deaf people here dont work in manual jobs that often in germany, they usually do some desk job in a "normal company"
04:49 AM Loetmichel: same for quadripledges (spelling?)
04:49 AM Loetmichel: (the ones with a broken spine or some other reason to sit in a wheelchair)
04:50 AM XXCoder: i would love a desk job, but i probably will fall asleep
04:51 AM Loetmichel: those "disabled workshops" are operated by non-profit NGOs, and USUALLY have more people with learning disabilities or the likes than bodily harm.
04:51 AM XXCoder: its pretty cool really
04:52 AM XXCoder: we are still figuring what my nephew could do
04:52 AM XXCoder: hes sensorory austic
04:52 AM Loetmichel: so he would be pretty good at a job that stays the same for long periods of time?
04:53 AM Loetmichel: isnt that an "on the spectrum" characteristig that they like routine and get irritated if something changes?
04:53 AM XXCoder: yeah i was thinking program debugging etc
04:54 AM jthornton: morning
05:00 AM Tom_itx: morning
05:01 AM XXCoder: morning
05:01 AM Tom_itx: jt, small typo on your gnipsel menu
05:01 AM XXCoder: jt i ran all day and probably made spoonful of swarf.
05:01 AM jthornton: what's that
05:01 AM XXCoder: chips heh
05:01 AM Tom_itx: the qt one
05:01 AM jthornton: no I meant the typo
05:02 AM Tom_itx: the Tutoria one?
05:02 AM XXCoder: lol both of us managed to not answer, tom
05:04 AM XXCoder: 70 parts and spoonful of chips lol
05:05 AM XXCoder: it just adds chamfers to 2 holes on both sidea
05:05 AM jthornton: what were you doing?
05:05 AM XXCoder: unfortunately I also had to use files to deburr parts
05:14 AM Tom_itx: jthornton, https://github.com/tecan/QPrint3D that's not the same as the QtPyvcp is it?
05:14 AM Tom_itx: no linuxcnc widgets i believe
05:15 AM Tom_itx: or are there linuxcnc widgets for c++
05:15 AM jthornton: no, that is something for 3d printing
05:15 AM Tom_itx: right
05:15 AM jthornton: no the linuxcnc widgets are in python
05:15 AM Tom_itx: but the environment is different
05:15 AM XXCoder: oh yeah whats linuxcnc status on python 2.0 vs 3.0?
05:17 AM jthornton: no one is working on anything that I know of except qtpyvcp and qtvcp
05:17 AM Tom_itx: he mentioned something about outputting gcode was why i looked at it
05:17 AM Tom_itx: ok
05:18 AM XXCoder: bit concerning as 2.0 ends in a year
05:18 AM jthornton: I wonder if he is reading the stl file or the g code from a slicer?
05:18 AM Tom_itx: i dunno
05:18 AM jthornton: will it be removed and no longer work?
05:18 AM XXCoder: i think simply unsupported
05:19 AM XXCoder: and later distros probably will remove. which isnt really a problem for linuxcnc as they roll their own distro modified
05:19 AM XXCoder: but no bugfixes essentally, and increasing outdated os
05:21 AM jthornton: I would not think that python 2 would have any bugs at this point
05:21 AM XXCoder: if it's bug-free it has bugs. if it has bugs it has 10 times more bugs
05:21 AM jthornton: no one is even doing that anymore
05:23 AM XXCoder: heh though i wonder how long python 2 would last with essentally no support
05:23 AM XXCoder: yo
05:25 AM jthornton: I guess at some point when it's not in a distro and not in a backport your screwed
05:25 AM XXCoder: yeah. linucnc would be blocked from using newer os
05:26 AM jthornton: I think linuxcnc has become so bloated with stuff that a stripped down version is needed and maybe that can be ported to python 3
05:27 AM XXCoder: some should be fobbed off into addon packages yeah
05:27 AM XXCoder: like image to gcode
05:27 AM XXCoder: also make addons that include some of your programs
05:28 AM jthornton: that's the approach that qtpyvcp has taken to be used with linuxcnc and not included in linuxcnc
06:03 AM The_Ball: skunkworks, finally making something usefull :) https://youtu.be/Kv09H5ngRZI
07:27 AM Tom_itx: jthornton, what would you strip out of it?
07:31 AM Tom_itx is now known as Tom_L
07:31 AM jthornton: everything but the core
07:31 AM Tom_L: maybe not a bad idea
07:31 AM jthornton: yea the python interface seems to be enough to run linuxcnc but the launcher now does additional stuff that is not needed or wanted in some cases
07:33 AM jthornton: you getting any rain yet?
07:36 AM Loetmichel: sooo, 10 fuse adapter sets done, 20 to go... maaan that takes ages on that small lathe... -> http://www.cyrom.org/palbum/main.php?g2_itemId=17352&g2_imageViewsIndex=1
08:47 AM jym: Loetmichel: Are you on crack? Seriously, fuse adapters???
08:47 AM Loetmichel: jym: nope.
08:48 AM jym: smh
08:48 AM Loetmichel: french military wants only to stock one type of fuse for their new PCs... if they want it i am willing to give it to them... not like the german VDE would object there ;)
08:49 AM jym: and new fuse folders wouldn't have been simpler?
08:49 AM Loetmichel: jym: fuse holders are an integral part of the mains filters... and they got ~50 PCs with two different series of filters
08:50 AM Loetmichel: one with the big fuses one with the small ones
08:50 AM Loetmichel: no option to change the holders
08:50 AM jym: no wonder they don't win any wars
08:52 AM jym: who thinks of these things?!
08:52 AM jym: nice job btw
09:01 AM gloops: you wouldnt want to be in full reverse in a tank with no spare fuse
09:03 AM jym: that's what foil is for, or .22 rounds
09:06 AM Loetmichel: jym: just get the powder out before using them as a fuse!
09:07 AM Loetmichel: otherwise that can have cathastropic effects i heard ;)
09:07 AM Loetmichel: gloops: the PCs are for office use. not in a tank
09:11 AM gloops: ahh
09:12 AM gloops: drilling hss tool steel, any experiences, advice? flat stock 3mm 4mm etc
09:15 AM jym: Loetmichel: .22LR bullet fuse... https://www.discovery.com/tv-shows/mythbusters/videos/gun-cartridge-fuse-minimyth
09:18 AM Loetmichel: jym: thats why i said "i heard"
09:18 AM Loetmichel: because i have no gun license and hence no ammo to test it
09:19 AM jym: Loetmichel: I was just saying, no need to remove the powder ;)
09:19 AM Loetmichel: the link is region locked though
09:19 AM Loetmichel: so i cant watch it
09:19 AM jym: oh geeeze
09:22 AM jym: Loetmichel: 31m25s https://www.dailymotion.com/video/x6sy431
09:25 AM jym is now known as Jymmm
12:45 PM Tom_L: jthornton, yeah light rain all morning
12:47 PM Tom_L: Loetmichel that's what CNC lathes are for
12:48 PM gloops: best wat to go about drilling HSS tool steel, about 4mm holes, 4mm thick blank
12:48 PM gloops: best way
12:49 PM Tom_L: cobalt?
12:49 PM Tom_L: as long as it's not hardened yet
12:51 PM gloops: its new steel so not hardened i dont think, so cobalt bits, loads of coolant i guess
12:58 PM cpresser: spiral toolpath with a 3mm carbide endmill?
12:59 PM gloops: cpresser only got woodworking spindle, i cant get the speed low enough i dont think
12:59 PM cpresser: but that would still be pretty close to a plunge-cut. I never tried to cut hss myself, so no clue how feasible that is.
01:00 PM cpresser: ask somebody to do it :)
01:02 PM gloops: well a pair of blanks for my shaper is £16 , the steel itself is about £2 to buy, i can grind the profiles myself, its just the location holes
01:03 PM gloops: probably end up spending more than £16 on burnt out drill bits to get one pair lol
01:10 PM Tom_L: could be
01:10 PM Tom_L: i know we cut tungsten with carbide cutters
01:11 PM Tom_L: but if you can't slow the spindle down it would likely just get hot
01:19 PM gloops: ill have a go on the bench drill, i can slow that down, see what happens, a hacksaw doesnt touch this steel
01:20 PM gloops: was hoping more than anything that there was some magical diamond tipped bit or something that flies through
01:20 PM gloops: that i hadnt heard of
01:22 PM Connor_remote: okay so, with 5i25 how with g540 firmware, how do you swap the pins for step / dir? I'm using a BOB that has those swapped
01:23 PM syyl_: unhardened hss machines very unproblematic, gloops
01:31 PM gloops: syyl_ yeah in an engineering shop maybe
01:31 PM FinboySlick: gloops: Didn't you have a ghetto EDM drill?
01:31 PM gloops: im using an old walker turner drill with a bottle of coolant lol
01:32 PM syyl_: yeah, what else would you need?
01:32 PM gloops: not me FinboySlick..
01:32 PM syyl_: its just a toolsteel
01:32 PM syyl_: not unobtainium
01:32 PM syyl_: dont go crazy fast and dont rub the drill to death :D
01:32 PM gloops: syyl_ you think i should be ok then? ill try it tommorrow, too late now
01:32 PM syyl_: yes
01:32 PM gloops: if i keep pecking
01:32 PM syyl_: as long as its unhardened
01:32 PM gregcnc: Are you sure it's soft if a hacksaw won't cut it?
01:33 PM FinboySlick: Was going to say, if you can't hacksaw it, it's probably hardened.
01:33 PM gloops: a hacksaw does nothing, not quick anyway, you get well bored by the time theres any mark
01:33 PM gregcnc: it's not soft
01:34 PM FinboySlick: gregcnc: maybe it's a copper hacksaw blade ;)
01:34 PM FinboySlick: How big a hole do you need, gloops?
01:35 PM syyl_: if its hardened
01:35 PM gloops: 4mm
01:35 PM syyl_: well
01:35 PM syyl_: good luck
01:35 PM syyl_: you can drill it on a mill with carbide, but even thats a stretch
01:35 PM FinboySlick: gloops: Well, if you didn't have a ghetto EDM, you should make one ;)
01:36 PM gloops: i might need some non-hardened stuff to start with then, far as i know this isnt hardened, it wasnt stated
01:36 PM roycroft: an oxy-acetylene torch will do the job with ease, if not with finesse :)
01:36 PM gloops: needs to be round about, to locate the blade
01:37 PM gloops: https://www.ebay.co.uk/i/172424709901?
01:37 PM gloops: like that
01:38 PM FinboySlick: Dead tree carcasses, disgusting!
01:39 PM gloops: so yes you would buy the cutters you can get off the shelf, but if you want custom profiles..
01:39 PM FinboySlick: gloops: https://www.youtube.com/watch?v=fRPmSgfIJqY
01:40 PM gloops: impressive FinboySlick
01:40 PM FinboySlick: For 40 bucks, you can do twice as good! ;)
01:40 PM gloops: but making machines to make machines, seem to have been down that road before lol
01:42 PM gloops: suppose i ought to establish whether i can drill this stuff first
01:46 PM Elmo40: does EDM work with higher voltage or current? is is along the lines of welding when you have low voltage but high amps?
01:47 PM FinboySlick: Intuitively I would say higer voltage but I don't know for sure.
01:48 PM gloops: miss0r made a pretty advanced spark erosion machine
01:48 PM FinboySlick: Obviously if you want a round hole, it's a good idea to make the tool spin too.
01:48 PM gloops: yeah his rotated
01:49 PM FinboySlick: Elmo40: https://www.youtube.com/watch?v=rpHYBz7ToII has a pretty good explanation.
01:53 PM Loetmichel: Elmo40: EDM works by creating tiny sparks that generate cavitation in the cutting fluid and by that do "ablation"
01:53 PM FinboySlick: Would linuxcnc handle negative feed rate?
01:53 PM Loetmichel: its not melting the metal (not all of it)
01:53 PM FinboySlick: (where EDM needs to back off)
01:53 PM Loetmichel: FinboySlick: why should it? feedrate cant be negative, its a speed. there is no negative speed
01:54 PM FinboySlick: Loetmichel: To simplify the control loop as per that last video I linked.
01:54 PM Loetmichel: FinboySlick: juse the peck drilling command
01:55 PM Elmo40: only negative direction
01:55 PM Elmo40: i was thinking of that, too.
01:55 PM Loetmichel: just in very tiny steps and long dwell
01:55 PM FinboySlick: Loetmichel: That's not quite as effective though.
01:55 PM Loetmichel: also the edm only has to back off if you have no forced liquid thru the center of the EDM electrode
01:56 PM Loetmichel: so you get clean fluid in the gap every now and then
01:56 PM FinboySlick: Loetmichel: Well, that Applied Science guy seemed to imply that both are pretty useful.
01:57 PM Loetmichel: FinboySlick: i made a few (manual) drill press EDMs for gettting broken taps out of engine blocks
01:58 PM Loetmichel: ciruit for the edm is dead simple
01:58 PM Loetmichel: and the feed down you can connect to the current signal
01:58 PM Loetmichel: more current= slower feed
01:59 PM Loetmichel: and every now and then pull out and go back in to get fresh fluid into the hole
02:00 PM Loetmichel: works pretty well, even with a simple drill press on looow rpm and regulating the current by hand on the quill lever
02:01 PM syyl_: inventing new ways of machining, interesting
02:01 PM syyl_: edm has to constantly back off
02:01 PM syyl_: to be anything more than a gloryfied spotwelder
02:01 PM FinboySlick: Hence my question about 'negative feed'.
02:02 PM Loetmichel: syyl_: nope. it has to maintain a certain gap
02:03 PM Loetmichel: not backing up
02:03 PM rmu: electric bell is a simple but effective construction
02:03 PM rmu: can EDM holes into razor blades *g*
02:04 PM rmu: https://youtu.be/uUN4_-xp1Wc?t=18
02:05 PM FinboySlick: rmu: Used indicator as a work clamp. Lost all credibility.
02:05 PM FinboySlick: I mean micrometer.
02:06 PM syyl_: uhm yes, miss-formulated
02:07 PM syyl_: it has to maintain the gap
02:07 PM syyl_: and if you want to do any deep-hole work retraction or flushing trough the electrode
02:08 PM FinboySlick: Could it be that the ability to 'feed negatively' has to do with instances where your feed step is a bit too large to maintain an ideal gap?
02:10 PM FinboySlick: Eg: current too low, feed, current too high.
02:10 PM FinboySlick: Then you'd be 'pecking' by one feed step until the situation is resolved.
02:11 PM Loetmichel: then your feed resolution is to low
02:12 PM Loetmichel: but as syyl said: if you do deep holes and have no flushing thru the (hollow) electrode you need to "peck drill" or retract every once in a while to get fresh fluid into the hole anyways
02:12 PM Loetmichel: but that is not "negative feed" thats simply peck drilling
02:12 PM Loetmichel: "pull out, move back into the old depth, start the current/feed regulation again
02:12 PM FinboySlick: In the applied laboratory video he's not talking about peck drilling though.
02:12 PM Loetmichel: it DOESENT work without a current feedback
02:13 PM Loetmichel: been there tried that
02:13 PM Loetmichel: a fixed feedrate dosent work if you have the SLIGHTEST flucuation in material or electrode or mains voltage
02:16 PM FinboySlick: I imagine he used that negative feed to correct for those then.
02:17 PM Loetmichel: the way i came up with was to use an aplifier to get the current signal to 0-5V (5V for optimal current) and send that to an analog input into LinuxCNC, use that input "inverted" as a feed override
02:18 PM Loetmichel: than set a peck drilling command with a low but reasonable feedrate and 0,2mm steps between pecks, and let it start
02:19 PM Loetmichel: it moves down at programmed feedrate until the current rises , then slows down to zero feedrate at 5V current signal... every 0,2mm it "pecks" and starts the cycle again
02:19 PM Loetmichel: s/linuxCNC/EMC2
02:19 PM Loetmichel: it was some time ago ;)
02:36 PM Elmo40: is there a way to make custom canned cycles? with movements controlled by different inputs ?
02:37 PM Elmo40: such as a value of resistance or change in pressure ?
02:59 PM Deejay: hi
03:56 PM Deejay: gn8
04:13 PM andypugh: Loetmichel: Actually LinuxCNC can do reverse run, controled by negative feedrates.
04:13 PM andypugh: We even had it mereged into Master for a while. I am not sure what the current status is
04:13 PM Loetmichel: andypugh: i never said it coudlnt
04:13 PM Loetmichel: i said "why should it?"
04:13 PM andypugh: rene_dev_: ? What happened there?
04:14 PM andypugh: Loetmichel: It could be handy for backing out a wire EDM if the part distorts enough to short the wire.
04:15 PM Loetmichel: yeah, less work than cutting the wire and restarting
04:15 PM andypugh: Or for plunge-edm to allow debris shorting the gap to fall out
04:15 PM Loetmichel: that should never happen.
04:16 PM Loetmichel: but if it does you have to stop and clean anyways
04:18 PM Loetmichel: but if that is the case the setup i described above would work even better
04:19 PM Loetmichel: you just have to amplify the current to 2.5V for optimum current and if it goes above the feed override gets negative
04:20 PM Loetmichel: the "peck drill cycle" is still needed for cleaning the cutting fluid in the plunge edm hole
05:05 PM skunkworks: so.... Spindle current -> pid -> adaptive feed... Thoughts?
05:05 PM skunkworks: *spindle power
05:14 PM _unreal_: do any of you use servos?
05:28 PM skunkworks: _unreal_: yes
05:32 PM Lcvette: skunkworks: interesting
05:33 PM Lcvette: that would certainly dumb down the feedrate learning curve some
05:34 PM Lcvette: at least on machines that can handle the spindle loading to that point
05:34 PM Lcvette: lol
05:35 PM Lcvette: but i suppose if you have control of the parameters for spindle current vs feedrate you can dial in whatever you want
05:35 PM Lcvette: sunds pretty cool
05:35 PM Lcvette: sounds*
05:37 PM Tom_L: Lcvette did you see his video on it?
05:37 PM Lcvette: no
05:37 PM Lcvette: where?
05:38 PM Lcvette: :D
05:38 PM Tom_L: i'm sure skunkworks would be glad to re'post it :)
05:38 PM Tom_L: i lost track of it
05:38 PM Lcvette: :(
05:38 PM Lcvette: https://www.youtube.com/watch?v=7LAtx8o48jI
05:38 PM Lcvette: ?
05:39 PM Tom_L: looks like it yes
05:55 PM skunkworks: I tried it on real hardware - it worked'ish... I was failing at tuning it though and ran out of time
05:55 PM skunkworks: Got a pretty good oscolation that I couldn't get rid of
05:56 PM Tom_L: something like that may need a 'self tuning' pid
05:57 PM Tom_L: don't ask me how they work...
06:24 PM andypugh: You can’t beat a goos osculation.
06:24 PM andypugh: (good osculation, I mean. Osculating a goose would be wierd, and often illegal)
06:40 PM Tom_L: https://www.controleng.com/articles/fundamentals-of-self-tuning-control/
07:12 PM gregcnc: skunkworks I think it might work in some materials, but in others you'll just be burning up the tool
07:13 PM _unreal_: skunkworks, what kind of servos?
07:14 PM _unreal_: I just got back from taking my daughter to the playground before the sun set
07:16 PM gregcnc: kuka robot sim in linuxcnc https://youtu.be/ByNznn5TDvQ
07:45 PM skunkworks: _unreal_: 40 ft-lb 8 brush servos, some brushless..
07:45 PM skunkworks: _unreal_: what is your quesion?
07:51 PM _unreal_: less of a question just wondering, I have smaller BLDC servo motors I'm looking at getting up and going
07:51 PM _unreal_: with built in encoders
07:51 PM _unreal_: I dont know what there working ratings are yet.
08:13 PM skunkworks: servos are cool
08:13 PM skunkworks: I like servos..
08:13 PM skunkworks: ;)
08:14 PM Elmo40: indeed they are
09:13 PM flyback is now known as phlyback
09:13 PM phlyback is now known as flyback
09:37 PM _unreal_: a3936sedt are the motor controllers I have that "" drive "" the servos I have
09:37 PM _unreal_: but they dont deal with the encoder
09:37 PM _unreal_: So I'm trying to figure out how to deal with them
09:37 PM _unreal_: I also dont know the working specs of the motors :/
09:38 PM _unreal_: torque etc...
11:36 PM net|: https://raw.githubusercontent.com/tecan/QPrint3D/master/screenshot.png