#linuxcnc Logs

Oct 14 2025

01:09 AM lcnc-relay: <meisterdippel@> moin
01:40 AM Deejay: moin
04:23 AM Tom_L: morning
07:38 AM JT-Mobile: must have had a power outage at home I don't see my other self's on line
08:28 AM lcnc-relay: <vibram@> Do you know if it's possible to monitor the actual velocity of one axis? I have a linear encoder on the axis wired with ethercat device. I have the number of pulses available but I don't have velocity information
08:32 AM Tom_L: may depend on the GUI you are using
08:34 AM Tom_L: https://www.gnipsel.com/linuxcnc/flexgui/labels.html#velocity-labels
08:34 AM lcnc-relay: <skunkworks8841@> Tom_L: I say this all the time
08:34 AM Tom_L: flexgui can
08:34 AM Tom_L: good morning sir
08:35 AM Tom_L: i reference most of my comments with axis or flexgui since that's what i'm familiar with
08:35 AM lcnc-relay: <vibram@> Tom_L: I'm in the setup phase so I can use axis or gmoccapy
08:35 AM Tom_L: most generally flexgui can
08:35 AM lcnc-relay: <vibram@> I don't care at this ooint
08:36 AM lcnc-relay: <vibram@> In Hal show I checked the motion current vel but this is the theoretical value
08:36 AM Tom_L: vibram if i were you i'd start with axis to get things working
08:36 AM lcnc-relay: <vibram@> Tom_L: Yup this is where I'm
08:36 AM lcnc-relay: <vibram@> I'm tuning the pid of my servo but struggling
08:36 AM Tom_L: i know jt does a calculated velocity of either 2 or 3 axis
08:37 AM lcnc-relay: <vibram@> 1 axis is fine for me now
08:37 AM lcnc-relay: <vibram@> But I don't know how to get it in Hal show. I don't need to modify the display, this is only for config purpose
08:38 AM Tom_L: axis_n_velocity_lb
08:38 AM Tom_L: n being 0-8?
08:38 AM Tom_L: ^^ flexgui
08:39 AM Tom_L: Note
08:39 AM Tom_L: The Axis velocity label only reports back jogging speed; use the joint velocity label for linear speed.
08:39 AM Tom_L: joint_vel_sec_n_lb ?
08:40 AM lcnc-relay: <vibram@> Let me checked again because I already had a look before 😒
08:52 AM lcnc-relay: <vibram@> I have the joint.0.vel-cmd
08:52 AM lcnc-relay: <vibram@> But this should not help me
08:52 AM Tom_L: the ones i'm showing are from flex
08:52 AM lcnc-relay: <vibram@> I don't know flex unfortunately, and not sure to switch now 🙁
08:52 AM Tom_L: https://www.gnipsel.com/linuxcnc/flexgui/index.html
08:52 AM Tom_L: my example has alot of 'test' elements included
08:53 AM Tom_L: although i'm not sure i've ever done an actual install. i've tested every aspect of it though
08:55 AM Tom_L: i run both axis and flex
08:56 AM Tom_L: and you can also run 2 separate configs
08:56 AM Tom_L: i do that and for the most part the configs are identical save the GUI stuff
09:09 AM lcnc-relay: <vibram@> OK thank you, I will have a lool
09:10 AM lcnc-relay: <vibram@> Gemini told me that I can compute with the scale feedback and ddt function, I will try tonight
10:06 AM pcw_at_home: There generally no point in using DDT to calculate velocity as the PID component already does this internally
10:07 AM pcw_at_home: There is
10:28 AM lcnc-relay: <vibram@> I will only use to roughly tune the pid. Someone gave me the following procedure to tune the pid:
10:28 AM lcnc-relay: 1. Increase max ferror
10:28 AM lcnc-relay: 2. Try to set ff1 with matching the velocity command with current velocity
10:28 AM lcnc-relay: 3. Adjust P and reduce the ferror
10:43 AM lcnc-relay: <vibram@> So the calculation is only for display
11:43 AM -!- #linuxcnc mode set to +v by ChanServ
12:06 PM xxcoder: interesting. https://www.tomshardware.com/software/windows/ditching-windows-10-heres-how-i-installed-windows-11-removed-ai-and-stripped-out-unnecessary-options-using-flyoobe
12:06 PM xxcoder: a way to install win 11 without ai crud and some stuff
12:35 PM mrec_: does anyone have experience with C and AI? openai gave me pretty smart answers. eg. when having a buffer and needing to add a trailing 0. (pseudo code): size=1024, char buf[size]; char old=buf[size]; buf[size]=0; do something with the string buf[size]=old;
12:36 PM mrec_: but I have had good experience with OpenGL based applications, and Qt boilerplate code
12:37 PM mrec_: state based pointer and buffer handling usually failed at some point
01:05 PM jpa-: mrec_: my experience is that it is garbage for anything complex, but ok for getting simple stuff done in a programming language you are unfamiliar with
01:07 PM mrec_: scurve calculation is also somewhat usable, I'm not completely done with it yet since I'm busy with something else
01:08 PM mrec_: deepseek was worst so far, but I was really surprised about what I got with opengl
01:23 PM rmu: those AI generated websites poisoning bona fide search results are really evil
01:24 PM rmu: stuff like this https://www.pythontutorials.net/
01:25 PM rmu: https://www.pythontutorials.net/blog/micropython-bluepill/
01:25 PM mrec_: python boilerplate code is fine
01:26 PM mrec_: overall I feel like I'm up to 10x faster with it (C / C++) and I can just focus on the core algorithms and application design
01:26 PM mrec_: writing testunits, documentation is also much easier now
01:27 PM rmu: you have to check it nonetheless if its not some hallucinated non-sense like the micropython on stm32f103
01:38 PM jpa-: mrec_: 10x more is not much if you are still at the level where you need to ask AI about putting a trailing 0.. it may actually be slowing your learning process, even if it is speeding up getting things immediately done
01:41 PM mrec_: jpa-: no I was doing a patch and it recommended fixing the bug that way
01:42 PM mrec_: of course it is ridiculous
01:43 PM lcnc-relay: <xxcoder@> test
01:43 PM mrec_: I found a bug in a system library, going the lazy way I asked AI to write a patch for it. The output was kind of blown up a lot finally I wrote it myself and still let AI double check it
01:43 PM xxcoder: nice. bridge be bridging now
01:44 PM mrec_: double checking the answer was I like your patch
01:44 PM mrec_: but still a few modifications which I did not agree with eg. the trailing 0
01:46 PM Guest67: looking for some help on ini vs hal files
01:46 PM mrec_: the testapp for the stm32 trapezoidal motion controller I decided to build with Qt, the boilerplate code from AI was absolutely okay
02:08 PM roycroft: yay
02:08 PM roycroft: my taxes are finally done and mailed off
02:09 PM roycroft: with a whole day to spare before the deadline
02:09 PM xxcoder: hy guest
02:09 PM xxcoder: *hey
02:09 PM xxcoder: probably need to be more detailed on your question :)
02:10 PM roycroft: a good general way to think about them is that an ini file is kind of like an include file for assigning values to variables
02:21 PM -!- #linuxcnc mode set to +v by ChanServ
02:23 PM mrec_: jpa-: do you use AI for programming support?
02:23 PM mrec_: syntax and small samples / boilerplate code is just perfect with it I feel
02:25 PM JT-Mobile: programming G code or Python or ?
02:26 PM mrec_: everything else than G-Code :-)
02:27 PM JT-Mobile: I find goggle ai is a good help for python syntax
02:53 PM JT-Mobile: hmm Flex GUI is starting to get a lot of users
02:58 PM Tom_L: really?
02:58 PM JT-Mobile: yup, more and more are using it
03:00 PM Tom_L: you can tell from the forum or some other way?
03:01 PM JT-Mobile: the forum is starting to have more and more users asking questions
03:01 PM Tom_L: super
03:01 PM JT-Mobile: I figure 1% ask and 99% don't ask
03:02 PM JT-Mobile: so chris said picking a gui was subjective and i reminded him that all the emc qtpy vcp are using an out of date qt
03:03 PM JT-Mobile: I'm the only one that has a current pyqt6 gui
03:03 PM Tom_L: and his reply?
03:04 PM JT-Mobile: none
03:04 PM xxcoder: honestly not a surpise. from what I see, flex gui looks amazing, and very customizable.
03:04 PM rdtsc-w: python for the win
03:05 PM JT-Mobile: the last one is doing a waterjet
03:06 PM JT-Mobile: https://forum.linuxcnc.org/41-guis/57433-flexgui-custom-buttons
03:06 PM Tom_L: well we've (you've) given it our best efforts
03:06 PM JT-Mobile: we both have
03:06 PM Tom_L: i just break sh*t
03:06 PM Tom_L: anyway, headin out the door again
03:07 PM JT-Mobile: testing and breaking is important
03:07 PM JT-Mobile: nap time here
03:07 PM rdtsc-w: say any idea how small of a hole a typical waterjet cuts?
03:07 PM Tom_L: hope their steaks are good. we quit going to the one here
03:07 PM Tom_L: smaller portions etc
03:10 PM JT-Mobile: I hope the steak is good
03:11 PM JT-Mobile: I had some parts for the can am spyder cut on laser
03:17 PM rdtsc-w: parts for articulating arm were fiber-laser; very clean, very accurate, almost zero distortion
03:22 PM JT-Mobile: yup those parts I had made were really good with no burrs or out of shape holes
03:29 PM lcnc-relay: <skunkworks8841@> In case anyone tries it - I guess you cannot run a pcie printer port on the rpi5 because it doesn't support io bar space mapping
03:41 PM lcnc-relay: <lucid.nonsense@> Could you do one off fpio?
03:41 PM lcnc-relay: <lucid.nonsense@> Gpio
03:42 PM lcnc-relay: <skunkworks8841@> the gpio works ok.. It is just very slow it seems
03:43 PM lcnc-relay: <skunkworks8841@> and really it doesn't matter - I was just goofing around seeing if a pcie printer port would be faster.
03:43 PM lcnc-relay: <skunkworks8841@> (and I would go mesa in the real world - the rpi5 runs it awesome)
03:44 PM lcnc-relay: <lucid.nonsense@> Did you enable pcie gen3?
03:44 PM lcnc-relay: <lucid.nonsense@> Know it's not officially supported, but maybe has more functionaliry
03:45 PM lcnc-relay: <lucid.nonsense@> Seems to run fine for nvme on my cm5
03:46 PM lcnc-relay: <lucid.nonsense@> No flappy pcie cable involved though
04:15 PM lcnc-relay: <allaboutjake@> Lcvette: Just curious, Was there a problem with the Debs in the end or should I keep trying to debug my probe basic issue on my own?
04:16 PM _unreal_: slowly migrating to this new linux system
04:16 PM _unreal_: the winpc is.... off line
04:17 PM _unreal_: I'm expecting MS to do some BS where they have an update force "fail" win10
04:18 PM _unreal_: I dont know if it is the OS but I'm getting some quiet buzzing in the speakers
04:19 PM _unreal_: I never plugged speakers in when it had win7
05:07 PM -!- #linuxcnc mode set to +v by ChanServ
05:42 PM -!- #linuxcnc mode set to +v by ChanServ
06:36 PM xxcoder: apparently ai rice cookers exist now
06:36 PM * xxcoder laughs with rice cooker that runs on pure power and physics
06:37 PM xxcoder: its thanks to tech connections alec that I know of, and own such rice cooker.
06:41 PM rdtsc: Headlines 2027: world taken over by AI rice cookers - all your base are belong to us!
06:50 PM xxcoder: https://www.youtube.com/watch?v=RSTNhvDGbYI he shows the exact brand and model I had gotten
09:06 PM roycroft: wow - i'm really rich, and i never knew it
09:07 PM roycroft: i just updated the metatada for all my books that are cataloged and shelved (most of them), and when i run stats on the collection it says the retail value of the collection is $550,011,496.68
09:08 PM roycroft: i'm thinking that if i sell half my books, i'll still have a nice collection, and an extra $225 million in my pocket (before taxes)
09:09 PM roycroft: there must be that one book that's worth a lot of money
09:09 PM * roycroft searches for it
09:15 PM Tom_L: ppl still buy books?
09:16 PM Tom_L: you'd really be if you found your code for the bitcoin
09:17 PM roycroft: i buy several books/month
09:17 PM roycroft: and i think i'm a person
09:17 PM roycroft: and yes, finding that pass phrase for my bitcoin would be really helpful
09:17 PM roycroft: i've not given up on finding it, but i'm also not holding out hope that i'll find it
09:18 PM roycroft: as i sort through more and more boxes of stuff i haven't looked at for years, both hope fades more and more that i'll find it and i get closer and closer to finding it
09:18 PM roycroft: i'm not a fatalist, so i don't just trust that fate will determine which is correct
09:19 PM roycroft: i figure i'll just find out one way or the other some day
09:19 PM roycroft: or maybe i'll die before i've gone through all my stuff, and i'll never know
09:26 PM _unreal_: fewwwww just finished typing up the working plan for a client
09:27 PM _unreal_: 3 pages, 2k words
09:46 PM t4nk_fn: half of them spelt wrong :b
09:47 PM t4nk_fn: ly.
09:52 PM lcnc-relay: <tom_cell@> test
09:55 PM xxcoder: hey cell of tom
09:56 PM Tom_L: can you change your nick?
09:56 PM xxcoder: bridge can only see username unfortunately
09:56 PM Tom_L: i had that from a very long time back
09:56 PM xxcoder: but you can change username also if needed
09:56 PM Tom_L: in discord...
09:56 PM Tom_L: how?
09:57 PM xxcoder: click gear icon your first page have "edit" for display name and username
09:57 PM xxcoder: on web client its by your nick at left bottom
09:57 PM xxcoder: no idea on phone app
09:57 PM Tom_L: i can set it on the pc
09:59 PM Tom_L: i never could get matrix to work right so i never bothered with it
09:59 PM xxcoder: last time I tried, I didn't like matrix ui
09:59 PM xxcoder: probably changed but whatever im on two already dont need third
10:02 PM roycroft: does it have a red button and a blue button?
10:03 PM xxcoder: no