#linuxcnc Logs
Sep 25 2024
#linuxcnc Calendar
12:55 AM Deejay: moin
01:29 AM lcnc-relay: <meisterdippel> moin
03:00 AM lcnc-relay: <meisterdippel> EPSON PS3 AS10
03:00 AM lcnc-relay: <meisterdippel> how can i find out which motors are installed there ?
03:44 AM -!- #linuxcnc mode set to +v by ChanServ
03:52 AM lcnc-relay: <vibram> you can send one to geneva 😄
04:11 AM lcnc-relay: <meisterdippel> for the right price, i can bring one to you 🙂
04:13 AM lcnc-relay: <meisterdippel> i can get one for ~ 450Euro, but without control-unit and no guarantee
04:13 AM lcnc-relay: <meisterdippel> I'm really thinking about it
04:19 AM JT-Cave: morning
04:22 AM lcnc-relay: <vibram> hummm unfortunately I have no use
04:22 AM lcnc-relay: <vibram> and retrofit would be very expensive I think
04:23 AM lcnc-relay: <vibram> which driver do you put? the open source one I can't remember the name but very complicated to source?
04:29 AM lcnc-relay: <meisterdippel> I have also no use, only playing around 🙂 i don't know what i can use as driver for the AC-Servos, but for the control i will use my LinuxCNC-Rio FPGA boards.
04:39 AM Tom_L: morning
04:41 AM Tom_L: JT-Cave, http://tom-itx.no-ip.biz:443/~tom-itx/irc/logs/%23linuxcnc/2024-09-24.html
04:41 AM Tom_L: 07:47 PM
04:43 AM JT-Cave: seems to be missing some info unless the keyboard interrupt is what killed it
04:44 AM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/flex_error.png
04:44 AM Tom_L: takes around 10 min to load unless it error'd out
04:45 AM Tom_L: if it loaded, it would hang with mdi commands etc
04:46 AM Tom_L: axis loads normally
04:46 AM Tom_L: same config
04:47 AM JT-Cave: touch probe 2 loads in a second here
04:48 AM Tom_L: sim
04:48 AM Tom_L: so does mine
04:51 AM Tom_L: i can't seem to find anything here but i'm still looing in the config
04:53 AM Tom_L: it's odd since i've been using that config for a while and this just started a few days ago and it's getting worse
04:53 AM Tom_L: tried to eliminate anything to do with pyvcp and the side pannel
04:54 AM Tom_L: since you can't use that
04:57 AM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch-probe2/config/
04:57 AM Tom_L: there's the config
05:04 AM Tom_L: been running that config for quite a while, i just tried to eliminate anything postgui.hal from it
05:48 AM JT-Cave: https://pasteboard.co/fuf09O2vZfWR.png
05:48 AM JT-Cave: changed the hal file to flex.hal and it runs fine after copying the resoruce.py and qss files
05:48 AM JT-Cave: note: the additions to your ui file :)
05:51 AM lcnc-relay: <captainhindsight_.> https://jauriarts.org/_matrix/media/v3/download/jauriarts.org/MePbSfavYkBmAThwBWwJcTCb/IMG_20240925_050737.jpg
05:52 AM lcnc-relay: <captainhindsight_.> meisterdippel maybe analog servo amps
06:00 AM lcnc-relay: <meisterdippel> yes, but which, or maybe self designed 🙂
06:49 AM rdtsc-wk: try to find a manual for the whole Epson robot
06:51 AM rdtsc-wk: https://epson.com/Support/Robots/sh/s6
07:29 AM rdtsc-wk: alas, meister left; probably dreaming of dirt-cheap robots with 400n-m servos
07:34 AM Tom_L: JT-Cave, the flex.hal content the same?
07:40 AM JT-Cave: the same as?
07:41 AM Tom_L: same as what i posted
07:41 AM Tom_L: i didn't see a flex.hal file
07:42 AM JT-Cave: I didn't see where you posted a flex.hal file
07:42 AM Tom_L: i didn't
07:42 AM JT-Cave: it's in xyz along with sim hardware hal
07:43 AM JT-Cave: https://pasteboard.co/P0Y5aVefnxOk.png
07:43 AM Tom_L: touch-probe2.hal is my primary hal file
07:43 AM Tom_L: is flex.hal required?
07:43 AM JT-Cave: yes, but I can't use it because it's for hardware
07:43 AM Tom_L: this isn't for a sim
07:43 AM JT-Cave: flex.hal is just a sim hal
07:44 AM JT-Cave: I can only test for a sim not the 7i90
07:44 AM JT-Cave: so I can't use your hal files
07:44 AM Tom_L: i have a sim too and it comes right up
07:44 AM JT-Cave: same here
07:44 AM Tom_L: i just can't figure out what's wrong with the live one
07:46 AM JT-Cave: make a new hal file with only the minimum hal to start up and copy chunks back in till it breaks
07:48 AM JT-Cave: [PYTHON]
07:48 AM JT-Cave: TOPLEVEL=./toplevel.py
07:48 AM JT-Cave: PATH_APPEND= ./
07:49 AM JT-Cave: what does that do?
07:50 AM JT-Cave: loadrt estop_latch is not used in your hal files...
07:52 AM * JT-Cave starts his chicken day
07:52 AM Tom_L: i think the estop latch may be for the pendant
07:52 AM Tom_L: i'll have to look it over
07:53 AM Tom_L: i have the python like that
08:37 AM lcnc-relay: <vibram> https://github.com/rene-dev/stmbl
08:37 AM lcnc-relay: <vibram> i was refering to this
08:37 AM lcnc-relay: <vibram> but 1 or 2 years ago it was difficult to source a complete driver
08:38 AM lcnc-relay: <vibram> and quite complicated to configure to me
08:49 AM -!- #linuxcnc mode set to +v by ChanServ
08:49 AM Tom_dev: JT-Cave, http://paste.debian.net/1330385/
08:49 AM Tom_L: been stripping things out of the hal file
08:50 AM Tom_L: until it tried to start without missing components
09:05 AM kworm: Would anyone happen to have a working example of what needs to be done on the linux interface/routing side for Ethercat EoE to work?
09:43 AM JT-Cave: Tom_L, interesting, I'll test some
09:44 AM JT-Cave: Tom_L, looks like you have a corrupt mdi_history.txt can you post that?
09:54 AM -!- #linuxcnc mode set to +v by ChanServ
09:54 AM JT-Shop: don't fix the file until you post it so I can see what caused the issue and detect it in code
09:57 AM rdtsc-wk: @vibram interesting, a STM32 drive. Comment there shows link to newer fork. But these would be a whole can of worms... better have lots of experience with such and lots of free time. Far easier to just buy a name-brand.
11:05 AM -!- #linuxcnc mode set to +v by ChanServ
11:05 AM Tom_L: JT-Shop, definitely something odd about it
11:06 AM Tom_L: i will post the file instead of the content
11:06 AM Tom_L: says it's 447Mb
11:09 AM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch-probe2/config/
11:09 AM Tom_L: looks like it's double saving or something
11:10 AM Tom_L: it would take alot of mdi commands in my lifetime to get that big a text file
11:10 AM Tom_L: i _do_ repeat some but not that many
11:12 AM lcnc-relay: <Travis Farmer> according to mine, it is tripple saving some commands
11:13 AM Tom_dev: i'll post the one from my mill file too, it has some that repeat 20-30 times
11:13 AM lcnc-relay: <Travis Farmer> https://github.com/travis-farmer/linuxcnc/blob/master/configs/my_LinuxCNC_router/mdi_history.txt
11:14 AM Tom_L: JT-Shop, i posted 2
11:14 AM Tom_L: named them from where they came
11:15 AM Tom_L: as far as the config, i may have to start over on it
11:16 AM Tom_L: it was freezing up after an mdi command
11:17 AM Tom_L: some repeated as many as 40 ish times
11:18 AM Tom_L: 10449472 lines
11:18 AM Tom_L: in the mill one
11:19 AM Tom_L: 24977966 in the probe one
11:19 AM lcnc-relay: <Travis Farmer> mine looks like 3 sets of 3 repeats, now that i look at it..., on one or two runs. so my file is just the beginning of something
11:21 AM lcnc-relay: <Travis Farmer> almost like every time the files is opened, it copies itself? but my last 4 lines are not repeats
11:23 AM Tom_L: that might explain why it progressively got worse over time
11:24 AM lcnc-relay: <Travis Farmer> check the commit history for my file on github, each commit is a post run
11:27 AM Tom_L: JT-Shop, i copied the config files i posted you this morning (i botched mine up) back to the touch-probe2 folder
11:27 AM Tom_L: deleted the mdi-history file and linuxcnc comes right up
11:27 AM Tom_L: i did save a copy of it
11:28 AM Tom_L: interesting bug
11:29 AM Tom_L: i was ready to pull what little hair i have left out last night
11:30 AM lcnc-relay: <Travis Farmer> very interesting indeed
11:30 AM lcnc-relay: <Travis Farmer> lol
11:30 AM Tom_L: trying to finish up the probe code but this started up
11:30 AM Tom_L: anyway i think the config is back like it was for now
11:34 AM lcnc-relay: <Travis Farmer> need to run logrotate on the mdi file, i guess 😉
11:35 AM JT-Shop: Tom_L, pastebin that mdi history file
11:37 AM lcnc-relay: <Travis Farmer> gotta take a shower... back in a bit...
11:46 AM Tom_L: jt i did
11:47 AM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch-probe2/config/
11:47 AM Tom_L: i don't think pastebin would like a 400M file
11:48 AM JT-Cave: got it coming
11:48 AM JT-Cave: dang it's huge
11:48 AM Tom_L: that's why lcnc was loading so slow here last night
11:48 AM JT-Cave: I assume it's mdi_history_probe2.txt
11:49 AM Tom_L: yes
11:49 AM Tom_L: the other one is from my regular mill folder using flex off and on
11:49 AM JT-Cave: did you have that much mdi history? or is the probe buttons adding it?
11:49 AM Tom_L: i didn't have near that much
11:49 AM Tom_L: well i didn't scroll in the window to see
11:50 AM Tom_L: i can if i need to
11:50 AM Tom_L: i moved the file from the folder and flex came right up
11:51 AM Tom_L: i'm not sure lcnc would start up if i put it back in the folder to scroll in the mdi window
11:51 AM Tom_L: i know for sure i didn't have 24977966 mdi commands though
11:52 AM JT-Cave: yea that's nuts
11:52 AM Tom_L: i can test and see what the probe buttons do later this afternoon
11:52 AM JT-Cave: ok
11:53 AM JT-Cave: nap time here
11:53 AM Deejay: dream well ;)
11:53 AM Tom_L: lunch then i gotta go for a while
11:53 AM JT-Cave: ok
11:53 AM Tom_L: i'll start with an empty history file and probe a bit
11:55 AM Tom_L: i'm not positive but i don't think i did any probing in the main mill folder
11:55 AM Tom_L: it's the smaller of the 2 history files i posted
11:55 AM Tom_L: i didn't have the probe yet
11:57 AM Tom_L: not a dull moment...
12:22 PM lcnc-relay: <Travis Farmer> back, now that everybody has left 😉
12:34 PM roycroft: we have a few puffy clouds today, and a much cooler temparture
12:34 PM roycroft: temperature
12:35 PM roycroft: i don't think it will rain, but i realized that a saw that i need for my roofing material is in an outbuilding at the old house, so i have to drive down to eugene today to get it
12:35 PM roycroft: which i'm fine with other than all the time it will take
12:35 PM lcnc-relay: <roguish> only 66 here right now. yesterday, it was around 85
12:35 PM roycroft: the logistics of getting everything done during my trip there tomorrow are rather challenging, and if i get some of those chores done today tomorrow will be much easier
12:36 PM * roycroft is looking forward to being done moving so he can enjoy the new place instead of contantly running away from it
12:36 PM xxcoder: lol
12:37 PM roycroft: this weekend i won't be moving much though - it's dump and recycle weekend
12:38 PM roycroft: i probably have 10 pickup loads of stuff to dispose of
12:38 PM roycroft: if i'm lucky i'll be able to haul 3 loads/day
12:38 PM roycroft: maybe 4
12:38 PM roycroft: the recycle center and the dump are on winter hours now, though, so they close at 5
12:54 PM lcnc-relay: <Travis Farmer> JT: you still here, or napping? i "may" have found the bug, given my limited python skills
12:57 PM lcnc-relay: <Travis Farmer> according to: https://www.w3schools.com/python/python_file_write.asp
12:57 PM lcnc-relay: file writes in python need to be closed afterward, with a f.close()
12:57 PM lcnc-relay: i don't see that in any of your MDI file writes, but maybe i missed it.
01:05 PM lcnc-relay: <Travis Farmer> "action_save_as" needs a f.close() i think, too
01:07 PM lcnc-relay: <Travis Farmer> found this on stackoverflow: "Close is always necessary when dealing with files, it is not a good idea to leave open file handles all over the place. They will eventually be closed when the file object is garbage collected but you do not know when that will be and in the mean time you will be wasting system resources by holding to file handles you no longer need."
01:11 PM lcnc-relay: <Travis Farmer> my guess is the "garbage collection" when LCNC is closed, appends all the file writes together then, and makes the actual file write
01:16 PM lcnc-relay: <Travis Farmer> maybe i am right, maybe completely wrong... but it is the only thing that popped off in my head
01:30 PM JT-Shop: travis when you use with open('file.txt', 'w') as f: it closes the file automagicly
01:32 PM lcnc-relay: <Travis Farmer> hmmm, nothing else is standing out...
01:39 PM lcnc-relay: <Travis Farmer> yeah, i just read about the "with open...", now i don't know
01:55 PM JT-Shop: I wonder what happens when you Ctrl C while the with open loop is still running?
02:06 PM lcnc-relay: <Travis Farmer> i think that would require infinatly fine timing to test
02:06 PM JT-Shop: not really just a time delay in the loop
02:07 PM * JT-Shop heads to town
02:30 PM lcnc-relay: <Travis Farmer> Hmmm, it seems to be raining... was not expecting that until tomorrow...
03:09 PM roycroft: well some rain has arrived, in spite of the optimistic forecast
03:09 PM roycroft: that actually further justifies my trip to the old house today - driving in the rain is a lot better than roofing in the rain
03:10 PM roycroft: assuming the rain has not ended when i get off work
04:05 PM roycroft: well, there's way more than 1mm of rain, and it's coming down pretty steadily
04:12 PM -!- #linuxcnc mode set to +v by ChanServ
04:13 PM Tom_dev: the probe routines don't seem to be adding any entries into the mdi_history file
04:13 PM Tom_dev: and right now i have 3 mdi entries i added via mdi
04:15 PM lcnc-relay: <Travis Farmer> https://jauriarts.org/_matrix/media/v3/download/jauriarts.org/FkNfgJfpEXOffmxPhqtixVAh/Screenshot_2024-09-25_at_17-07-00_motionEye_Home_Assistant.png
04:15 PM lcnc-relay: <Travis Farmer> deer eating acorns off my lawn
04:16 PM lcnc-relay: <Travis Farmer> https://jauriarts.org/_matrix/media/v3/download/jauriarts.org/UbQVIKgciQYVqCHvlYvLSFeF/Screenshot_2024-09-25_at_17-09-19_motionEye_Home_Assistant.png
04:16 PM lcnc-relay: <Travis Farmer> bigger
04:18 PM lcnc-relay: <Travis Farmer> Tom_L: i wonder if it was a bug that solved itself without notice, and we just noticed the effect?
05:00 PM -!- #linuxcnc mode set to +v by ChanServ
05:00 PM JT-Shop: somehow I signed up for prime...
05:04 PM Tom_L: lucky you
05:05 PM Tom_L: the probe routines didn't add any entry in the mdi history file
05:05 PM Tom_L: i checked all but the tool setter
05:06 PM JT-Shop: I tried to duplicate the issue in python but could not
05:06 PM JT-Shop: I'll prob program a limit to the number of lines written to the mdi history file
05:07 PM Tom_L: an odd one for sure
05:07 PM JT-Shop: may check file size before trying to load it as well
05:07 PM Tom_L: i was gonna look at the top of the file and see if it was a point where it was 'full' and that triggered it
05:07 PM Tom_L: i dunno
05:08 PM Tom_L: turn the file into a fifo buffer
05:08 PM Tom_L: so if it fills, the top gets flushed off
05:08 PM Tom_L: just an idea
05:09 PM Tom_L: with the file cleared, my config runs fine
05:10 PM Tom_L: initial checks say it's not coming from the probe routines though
05:10 PM JT-Shop: yup I checked that too
05:10 PM Tom_L: after i ran thru them with an empty file i did a couple of mdi commands and that was all that was in the file
05:15 PM JT-Shop: another chap on the forum chimed in he is using flex on his industrial router
05:15 PM Tom_L: nice
05:16 PM Tom_L: i'd advise about the mdi file or at least encourage an update once you figure it out
05:17 PM JT-Shop: I have no clue how your mdi history file got that big... interestingly the lines all seemed to be g code
05:18 PM Tom_L: yeah me either
05:18 PM Tom_L: i've tested alot but my god.. my fingers would be wore to the bone
05:21 PM JT-Shop: it would decades to do that many mdi commands
05:23 PM JT-Shop: I honestly think we can just ignore that and move on
05:23 PM Tom_L: fine by me
05:23 PM Tom_L: unless it crops up again
05:24 PM Tom_L: i noticed last few days linuxcnc didn't seem to be responding normal
05:24 PM Tom_L: then last couple days it started taking like 10 min to load
05:25 PM Tom_L: now i need to run thru the metric subs and make sure they're ok now
05:25 PM Tom_L: i have some rapid moves in them but i'm thinking about making them all G1
05:26 PM Tom_L: any thoughts?
05:26 PM JT-Shop: you might have a metric config to test taht
05:26 PM JT-Shop: that
05:26 PM Tom_L: i have one i just have to watch carefully
05:27 PM Tom_L: they'll run under an inch config (don't tell anybody)
05:27 PM JT-Shop: if you convert the inch to metric in your config then the rapid speeds will be the same
05:27 PM Tom_L: i decided on going with 2 sets of files
05:27 PM Tom_L: and they're all ready but the final test
05:29 PM Tom_L: when i run them i convert the result to inch and they are spot on
05:30 PM Tom_L: anyway, i doubt i get anything done tonight on it
05:32 PM JT-Shop: I'm relaxing watching Leo build Tally Ho for the umteenth time
05:32 PM JT-Shop: and sipping on my vino
05:36 PM Tom_L: good call
05:36 PM Tom_L: at least i can relax knowing it's not my config
05:37 PM JT-Shop: I don't think it's flex either but who knows
05:38 PM Tom_L: i'll try to keep an eye on that file size
05:38 PM JT-Shop: good idea
05:40 PM JT-Shop: my corvette buddy is going to pick up a trailer load of c3 parts tomorrow... for him it's a sickness he said, I assume it's because he had a corvette shop in st louis
05:40 PM Tom_L: the axis set buttons need a little attention
05:40 PM Tom_L: heh maybe you should ride along
05:41 PM JT-Shop: attention from me or you?
05:46 PM -!- #linuxcnc mode set to +v by ChanServ
06:22 PM Tom_L: me
06:57 PM rdtsc_away is now known as rdtsc
07:29 PM rdtsc is now known as rdtsc_away
08:20 PM Tom_L: JT-Cave, PR #94
08:24 PM Tom_L: after testing mdi_history was 43k
08:24 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch-probe2/config/mdi_history.txt
08:24 PM Tom_L: in case you wanna look at it
08:24 PM Tom_L: none came from probing
11:05 PM lcnc-relay: <Travis Farmer> oh deer... i honestly thought that was solved... 😦