#linuxcnc Logs

Jun 15 2025

12:04 AM roycroft: who is that?
12:19 AM xxcoder: https://www.youtube.com/watch?v=w-SSGs8mYrk this guy
12:19 AM roycroft: oh
12:19 AM roycroft: i've seen those videos before but never watched the
12:19 AM roycroft: m
12:19 AM roycroft: i'll check it out when i have a chance
12:19 AM xxcoder: he makes a lot of random stuff using 3d printer, casting, and machining
12:19 AM xxcoder: that video is second part btw, he makes variant of same thing
02:06 AM Deejay: moin
04:50 AM lcnc-relay: <travis_farmer@> Morning
04:56 AM lcnc-relay: <sodakaustik@> mornin!
04:58 AM lcnc-relay: <sodakaustik@> Is a torque-controlled spindle a good idea for a lathe? 🤔
04:58 AM lcnc-relay: my spindle encoder only has 1600ppr, which tells me it might make the whole thing a bit jittery.
04:58 AM lcnc-relay: Its a blind vector vfd. Maybe i should mirror the encoder data to both the vfd and linuxcnc...
05:23 AM Tom_L: morning
08:34 AM lcnc-relay: <tgunr@> Happy Fathers day to the fathers.
08:42 AM lcnc-relay: <tgunr@> The docs says "LinuxCNC uses a parameter file to ensure persistence". I can't seem to find it, where is it typically located? Or, what I really want is to read in a file with various global values I want for different macros. I suppose I can read in a file with subroutine?
08:45 AM lcnc-relay: <tgunr@> xxcoder: Thats cool, always wanted to work with metal like that.
08:45 AM Tom_L: in the config directory.
08:45 AM Tom_L: .var file
08:46 AM Tom_L: if you want to create your own user parameters, you have to define them in the file first or they won't be persistent
08:46 AM lcnc-relay: <zincboy_ca_on@> See https://www.linuxcnc.org/docs/html//gcode/overview.html#sub:numbered-parameters
08:47 AM Tom_L: beat me to ti
08:47 AM Tom_L: it
08:48 AM Tom_L: 31-5000 - G-code user parameters. These parameters are global in the G code file, and available for general use. Volatile.
08:48 AM Tom_L: it says volatile but it doesn't say if you define them they are persistent
08:53 AM Tom_L: also that var file name can be defined in the ini
08:53 AM Tom_L: http://linuxcnc.org/docs/stable/html/config/ini-config.html#sub:ini:sec:rs274ngc
08:57 AM Tom_L: you can also pass parameters to a sub from a command line. i forget the variable range on that but it's in the lower numbers
08:58 AM Tom_L: 1-30
08:58 AM Tom_L: https://www.linuxcnc.org/docs/html//gcode/overview.html#sub:subroutine-parameters
09:10 AM lcnc-relay: <tgunr@> Tom_L: I assume you mean "/var" but I don't see anything interesting in my "/var/lib/bbctrl"
09:14 AM Tom_L: nope i mean /linuxcnc/configs/yourconfig/parameter.var
09:15 AM Tom_L: or linuxcnc.var rather
09:15 AM Tom_L: https://github.com/tlight9/flexgui/blob/master/examples/touch-probe3/linuxcnc.var
09:15 AM Tom_L: from my examples
09:16 AM Tom_L: i added the ones from #2000 to #2120
09:16 AM Tom_L: otherwise they would start at #5161 iirc
09:18 AM Tom_L: you're not running linuxcnc either right?
09:21 AM Tom_L: your controller may not support persistent parameters
09:21 AM Tom_L: https://buildbotics.com/manual-v2-0-0/
09:21 AM lcnc-relay: <off7788@> tgunr@: /home/xxx/linuxcnc/configs/xxxx/xxx.var
09:21 AM Tom_L: i'm not familiar with it
09:22 AM lcnc-relay: <tgunr@> the only user in this /home is /home/bbmc and nothing interesting
09:22 AM Tom_L: your controller may not support persistent parameters
09:22 AM lcnc-relay: <tgunr@> will look thru the sources
09:22 AM Tom_L: you're asking linuxcnc questions about a foreign control
09:23 AM lcnc-relay: <tgunr@> I'll send it back if so
09:23 AM Tom_L: lcnc is alot more flexible but comes with a slight learning curve
09:23 AM lcnc-relay: <tgunr@> or use its raspberry as a proper linuxcnc if possible
09:23 AM Tom_L: which rpi is it?
09:24 AM Tom_L: 4 & 5 seem to work good
09:24 AM Tom_L: anything less not so much
09:24 AM lcnc-relay: <tgunr@> it is supposedly based on linuxcnc, close but no cigar
09:24 AM lcnc-relay: <tgunr@> its a 4b i think
09:24 AM lcnc-relay: <sodakaustik@> I am using torque mode on the lathe spindle and the starts and stops are genuinely terrifying
09:24 AM Tom_L: should work
09:25 AM Tom_L: https://www.youtube.com/watch?v=g0OLIth_cOI
09:25 AM Tom_L: crap video but using a rpi4
09:27 AM Tom_L: you can't do much else with it while running lcnc though
09:27 AM Tom_L: the rpi5 would be a better fit. i'm getting one today to try
09:30 AM Tom_L: you can use mesa ethernet cards or spi specific for the rpi
09:52 AM rdtsc: sodakaustik, not surprising; allowing speed to do whatever it wants in favor for constant torque... once had a big drive try to spin up a one-ton motor in 0.1s due to being in constant-torque mode instead of constant velocity... nearly flipped the motor test cart it was bolted to!
10:31 AM lcnc-relay: <captainhindsight_.@> just FYI 32 bit RTAI is no longer supported so EOL
10:31 AM lcnc-relay: <captainhindsight_.@> End of Life for a few years now
10:32 AM lcnc-relay: <captainhindsight_.@> anyone that "needs" it can use an older distro
10:32 AM Tom_L: i thought so but he seemed pretty persistent about it
10:32 AM Tom_L: and that's what i told him
10:32 AM lcnc-relay: <captainhindsight_.@> yeah, i dodn't dig into his details for why , maybe he has old hardware
10:33 AM lcnc-relay: <captainhindsight_.@> dodn't/didn't
10:33 AM Tom_L: maybe trying to run some old software
10:33 AM Tom_L: i forget now
10:35 AM lcnc-relay: <captainhindsight_.@> anyway memleak are out of the loop here until fall
10:35 AM Tom_L: np
10:35 AM lcnc-relay: <captainhindsight_.@> memleak and I
10:35 AM lcnc-relay: <captainhindsight_.@> have to get my brain fixed
10:35 AM Tom_L: yup, focus on that for sure
10:36 AM Tom_L: my friend's wife had one and it drastically changed her mood
10:36 AM lcnc-relay: <captainhindsight_.@> if something does come up RTAI wise, meleak/ntulinux has and email for questions
10:37 AM Tom_L: i don't think many are using it anymore
10:37 AM Tom_L: maybe a handfull
10:37 AM lcnc-relay: <captainhindsight_.@> I'm mostly frustrated with the big miss by the radiologists for 18 months
10:38 AM Tom_L: i was talking to a doc as the mrs recently had a pacemaker put in. he said they are very limited by corporate as to their time management
10:39 AM lcnc-relay: <captainhindsight_.@> keep pushing your docs for answers if you're not feeling right
10:39 AM Tom_L: he was getting messages as he took the time to visit with us a bit
10:39 AM lcnc-relay: <captainhindsight_.@> I was getting the "it's in your head"
10:39 AM lcnc-relay: <captainhindsight_.@> yeah but mine is literal
10:40 AM lcnc-relay: <captainhindsight_.@> 🧠
10:41 AM lcnc-relay: <captainhindsight_.@> one surgeon actually said " I'll try to fit you in between trips outside the USA"
10:41 AM Tom_L: when the cardiologist came in he hadn't even looked at the scans while giving his diagnosis
10:42 AM lcnc-relay: <captainhindsight_.@> and then I heard him talking about Hawaii in the hallway for 10 minutes
10:42 AM Tom_L: i quit my family doctor because he was more interested in his plane than patients
10:43 AM lcnc-relay: <captainhindsight_.@> I changed hospitals. I found a group that specializes in my issue.
10:43 AM Tom_L: it's widely known among them
10:48 AM lcnc-relay: <captainhindsight_.@> most of the regs and users here have more skills than all the CNC "engineers" I have come across working for multi-billion $$$$ corporations
10:48 AM lcnc-relay: <captainhindsight_.@> and certainly are sharper
10:49 AM lcnc-relay: <skunkworks8841@> captainhindsight_.@ What is going on?
10:49 AM lcnc-relay: <captainhindsight_.@> broken brain
10:49 AM lcnc-relay: <captainhindsight_.@> just some hole drilling and resection
10:51 AM lcnc-relay: <skunkworks8841@> The problem that Billy Joel just had with his brain - my brother in law had a few years ago.. brain fluid wasn't draining..
10:51 AM lcnc-relay: <captainhindsight_.@> what I'm trying to avoid
10:52 AM lcnc-relay: <captainhindsight_.@> my case is more like Mark Ruffalos
10:53 AM lcnc-relay: <roguish> with anything medical, for yourself or anyone else, ya gotta agressively advocate and push.
10:53 AM lcnc-relay: <captainhindsight_.@> I thought it was side effects of my cancer treatments
10:53 AM lcnc-relay: <skunkworks8841@> Yikes.. good luck!
10:54 AM lcnc-relay: <captainhindsight_.@> what roguish said ^^^
10:55 AM lcnc-relay: <captainhindsight_.@> they almost missed my cancer as well 2 years ago
10:56 AM lcnc-relay: <roguish> as we all get older, we gotta really watch and listen to our bodies..... finding a good, caring GP is difficult, but essential
10:58 AM lcnc-relay: <captainhindsight_.@> it's also more difficult the farther you get from large cities
10:58 AM lcnc-relay: <roguish> that's probably true too
10:58 AM lcnc-relay: <captainhindsight_.@> I have choices
11:01 AM lcnc-relay: <roguish> i've been considering moving out of the Bay Area (too crowded and expensive) and having good senior health care is way high on my criteria list
11:11 AM lcnc-relay: <captainhindsight_.@> https://www.mayoclinic.org/locations
11:13 AM lcnc-relay: <captainhindsight_.@> https://health.usnews.com/health-care/best-hospitals/articles/best-hospitals-honor-roll-and-overview this list has been invaluable to be
11:17 AM lcnc-relay: <captainhindsight_.@> be/me 🙂
11:17 AM lcnc-relay: <roguish> intereesting list
11:17 AM lcnc-relay: <roguish> 2 of them here. Stanford, and UCSF
11:18 AM lcnc-relay: <travis_farmer@> man, this oppressive heat is not helping my severe depression...
11:19 AM lcnc-relay: <roguish> one where I'm thinking of. UCSD. list doesn't have Scripps in the San Diego area
11:20 AM lcnc-relay: <roguish> travis_farmer@: stay in the shade, and turn a fan on
11:21 AM lcnc-relay: <travis_farmer@> i am inside, and mother has the only fan pointed at her...
11:22 AM lcnc-relay: <captainhindsight_.@> 🌪️ best I can do from here Travis
11:24 AM lcnc-relay: <captainhindsight_.@> roguish000@ https://health.usnews.com/best-hospitals/area/san-diego-ca
11:25 AM lcnc-relay: <captainhindsight_.@> #6 in CA
11:38 AM roycroft: cook county hospital is on the list of best hospitals
11:39 AM roycroft: i wonder why that is - the last time i was there i only had to wait 6 hours in the er before someone talked to me
11:39 AM roycroft: it is not on the list, rather
12:47 PM lcnc-relay: <roguish> roycroft: when I need an ER or an Urgent Care center, I usually call around to find the shortest waiting list...
12:57 PM lcnc-relay: <sodakaustik@> does anyone know the bit depth of the pid calculation function
12:59 PM lcnc-relay: <skunkworks8841@> I assume it is 64 bit.. but I could be wrong..
01:00 PM lcnc-relay: <skunkworks8841@> Floats
01:14 PM lcnc-relay: <sodakaustik@> In getting something strange
01:14 PM lcnc-relay: <sodakaustik@> quantization on the pid output
01:23 PM pcw---home: quantization is expected because the encoder feedback is quantized
01:30 PM pcw---home: Imagine a Servo with a high resolution 17 bit encoder. With direct drive of a 5mm pitch ballscrew and a 10 micron error band, the full PID range is only about 8 bit resolution
01:39 PM roycroft: roguish: that's what happened when i broke my back
01:40 PM roycroft: i went to an urgent care facility, thinking the problem was my head, not my back, and they told me i'd need imaging, and they could not do that
01:40 PM roycroft: they checked the local hospitals and sent me to the one with the 10 minute wait
01:40 PM roycroft: which for me was actually zero minutes
01:40 PM roycroft: when i walked in i was in a wheelchair in seconds, with a brace around my neck, before i even gave them my insurance info
01:41 PM roycroft: and after the paperwork they wheeled me into the waiting area, and by the time the person that did that walked away a nurse was approaching to take me in for treatment
01:42 PM roycroft: i guess they thought it was a big deal
01:42 PM xxcoder: lawsuit avoidance
01:42 PM roycroft: i'd call it good triage
01:43 PM * roycroft just came up with a cunning plan for the new roof framing, and got the rear panel design for his minirouter controller started, and thinks that's enough cad work for the morning
01:44 PM roycroft: it's time to cut the controller cabinet pieces to final dimension and start machining them
01:44 PM roycroft: my power inlet is not arriving until tomorrow, so i won't work on the back panel until the part is in hand, but i can get most of the rest done today
01:48 PM lcnc-relay: <sodakaustik@> is it possible to only lowpass the D term of the pid loop?
01:49 PM lcnc-relay: <sodakaustik@> well, at least lowpass it differently.
01:50 PM pcw---home: You typically never want to delay D feedback
01:50 PM lcnc-relay: <sodakaustik@> works wonders in the drone world
01:51 PM pcw---home: Drones have very low bandwidth (a several hz at most)
01:53 PM pcw---home: If you are using Mesa hardware you can use the encoder velocity estimation to increase the velocity feedback resolution
01:54 PM lcnc-relay: <sodakaustik@> Im doing an orient move with a torque mode spindle
01:54 PM lcnc-relay: <sodakaustik@> D term is catastrophic because of my 1600ppr spindle encoder
01:54 PM lcnc-relay: <sodakaustik@> I am quite serious about the filtering
01:55 PM pcw---home: (by connecting the encoder velocity pin to the PID velocity feedback pin)
01:55 PM pcw---home: are you using the encoder velocity pin?
01:55 PM lcnc-relay: <sodakaustik@> fpv drones D term operates at about 120Hz and P/i are at around 450 btw
01:55 PM pcw---home: Yes but the "plant" has only a few hz resolution
01:56 PM lcnc-relay: <sodakaustik@> I am using the encoder position
01:56 PM pcw---home: Right, this gives very crunchy velocity feedback
01:57 PM lcnc-relay: <sodakaustik@> Yes, but it is the only way to get a correct position, right?
01:57 PM pcw---home: because the D term is calculated by D/DT of position
01:57 PM lcnc-relay: <sodakaustik@> I use Velocity PID in normal spindle operation
01:58 PM lcnc-relay: <sodakaustik@> which gives a more tolerable D term, which i still dont use because the benefit is zero. Instead i am ramping the spindle command in this case
01:59 PM pcw---home: is pid.N.feedback-deriv connected?
02:01 PM lcnc-relay: <sodakaustik@> good point, i dont think so, sadly i left the shop and i dont remember but i am fairly certain it isnt
02:03 PM pcw---home: D/DT of position at say 1000 Hz servo thread and slow speeds is going to be very noisy
02:04 PM pcw---home: the PID component uses D/DT of position for the D term if pid.N.feedback-deriv is not connected
02:05 PM pcw---home: ( very noisy because at low speeds you will get 1 count/0 count alternations )
02:06 PM pcw---home: using the encoder velocity estimation improves this a lot
02:07 PM pcw---home: (without delays that would compromise loop stability)
02:08 PM lcnc-relay: <sodakaustik@> makes sense
02:08 PM lcnc-relay: <sodakaustik@> is this the inclination between position pulses
02:08 PM lcnc-relay: <sodakaustik@> if we speak graphically
02:08 PM lcnc-relay: <sodakaustik@> sounds like it
02:17 PM pcw---home: So if you have a 1600 PPR encoder (6400 counts/turn) are you rotate slower than about 10 RPM you will get D/DTs of 1,1,1,1,0,1,1,1 or some such leading to noisy feedback
02:19 PM pcw---home: Velocity estimation uses counts/time between counts instead of counts/servo thread period so is much cleaner
03:28 PM lcnc-relay: <tgunr@> Tom_L: maybe he has a Beech Bonanza? Also know as "The Doctor Killer"
03:35 PM lcnc-relay: <sodakaustik@> 1600 counts per turn* its 400ppr 😭
03:35 PM lcnc-relay: <sodakaustik@> Ill connect this pin
03:35 PM Tom_L: or a G6?
03:58 PM lcnc-relay: <tgunr@> G38.2 is based on G90 & G91 right?
04:15 PM Tom_L: it will affect the outcome yes
04:29 PM -!- #linuxcnc mode set to +v by ChanServ