#linuxcnc Logs

Mar 11 2023

01:35 AM Deejay: moi
05:11 AM Tom_L: morning
05:21 AM -!- #linuxcnc mode set to +v by ChanServ
05:21 AM JT-Cave: morning
05:22 AM -!- #linuxcnc mode set to +v by ChanServ
05:24 AM JT-Cave: I see there is a python3-opengl update on debian 13 so I'm going to update and cross my fingers that axis works after
05:37 AM -!- #linuxcnc mode set to +v by ChanServ
05:58 AM -!- #linuxcnc mode set to +v by ChanServ
06:00 AM travis_farmer[m]: oh, now that it is morning and not evening... G'Morning ☕️
06:37 AM spray[m]: Can I get a recommendation please for optimal PC specs for running linuxcnc?
06:38 AM spray[m]: If I need an i7 or some old workstation is sufficient
07:16 AM -!- #linuxcnc mode set to +v by ChanServ
07:17 AM JT-Shop: after an apt dist-upgrade on d13 at least linuxcnc runs now
07:17 AM JT-Shop: my lan adapters are acting up now lol
08:02 AM -!- #linuxcnc mode set to +v by ChanServ
08:59 AM -!- #linuxcnc mode set to +v by ChanServ
09:32 AM -!- #linuxcnc mode set to +v by ChanServ
10:02 AM Rowow5457[m]: Anyone got a suggestion for a cheap working wifi adapter for Linux cnc?
10:02 AM Rowow5457[m]: I got a ac600 from TP-Link and it's not working
10:03 AM Rowow5457[m]: Tried 2 other adapters too. Not sure what I'm doing wrong the drivers are installed
10:05 AM * Rowow5457[m] uploaded an image: (685KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/gAjqqhKjCwqqVbroUmEdBiPE/PXL_20230311_160410455.jpg >
10:05 AM * Rowow5457[m] uploaded an image: (918KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/vnUqMfCGVeBxmNbevzuoqdDD/PXL_20230311_160420255.jpg >
10:06 AM Rowow5457[m]: Also finally got in the mesa controller yesterday
10:43 AM roguish[m]1: Rowow: it's best if you use the on-board nic for the Mesa board, and a secondary nic for the local network.
10:44 AM unterhaus_: he wants a wifi for his local network
10:45 AM unterhaus_: setting up a wifi adapter is the same for any linux, not really a lcnc question
10:45 AM roguish[m]1: yup.
10:45 AM unterhaus_: I suppose the first thing to do is look in dmesg to see if the wifi adapter is running
10:46 AM unterhaus_: what linux distribution is this on?
10:50 AM Rowow5457[m]: thats what im doing, currently using the onboard rj45 port for the mesa
10:50 AM Rowow5457[m]: looking to add wifi so I dont have to use usb drives all the time and look up tutorials/etc
10:51 AM roguish[m]1: Rowow: that's cool. I personnally prefer wired connection over wifi, but whatever floats your boat.
10:51 AM Rowow5457[m]: so it should just be plug and play? I was worried that linux cnc had something special or I had to configure it. or maybe I messed up doing the initial install
10:52 AM Rowow5457[m]: ill try this in a bit thanks. Ill double check the distro but I installed the USB installation drive last week so should be the latest.
10:55 AM roguish[m]1: Rowow: a nic should be pnp, but the dual setup for your wifi and mesa usually takes a bit of setup.
10:55 AM roguish[m]1: look here for some network setup help https://forum.linuxcnc.org/27-driver-boards
10:55 AM unterhaus_: wifi always has to be configured to your local network. They probably could hack into your network, but they might get in trouble
10:56 AM unterhaus_: windows used to connect to your neighbor's wifi if the signal was stronger, but only if it was open
10:56 AM roguish[m]1: https://forum.linuxcnc.org/27-driver-boards/35592-mesa-ethernet-setup-for-7i92-7i96-7i76e-and-other-ethernet-mesa-boards
11:10 AM Rowow5457[m]: thanks so much this helps a ton in leading me in the right direction!
11:23 AM unterhaus_: my garage door spring broke and the torsion tube looks horrible. I am trying tofigure out if 1" emt will work and the internet isn't helping
11:23 AM unterhaus_: it sure looks like 1" emt
11:26 AM Tom_L: conduit?
11:26 AM ZincBoy[CAON][m]: Torsion springs are about the last thing I would wing it on.
11:26 AM Tom_L: i doubt it
11:26 AM Tom_L: it's probably heavier than that
11:26 AM Tom_L: meh
11:27 AM Tom_L: you ratchet the torsion on them
11:28 AM ZincBoy[CAON][m]: Yes, but I mean if the tube you are ratcheting them on fails the you are in for a bad time.
11:29 AM Tom_L: mine are the other style and i've had them pop while the door was going up with me standing nearby
11:29 AM Tom_L: i've got the safety wire on all of them though
11:29 AM Tom_L: still makes a pretty good pop
11:30 AM Tom_L: those are color coded by tension
11:30 AM Tom_L: i dunno about torsion springs
11:38 AM genericum: I'm trying to write a gcode parser, whats the difference between G0 X10 Z10; Z0; and G0 X10 Z10 Z0; ? should I scan for line feeds only?
11:39 AM jpa-: is the ";" a literal semicolon or a newline?
11:39 AM ZincBoy[CAON][m]: I don't think the second one is valid.
11:39 AM jpa-: because ; is a comment character
11:39 AM unterhaus_: I think the semicolon makes it a comment
11:40 AM ZincBoy[CAON][m]: I don't think you can have two axis words on the same line. Behavior would be undefined.
11:40 AM jpa-: yeah, specifying axis value twice is not allowed by the standard - i think linuxcnc gives an error, many simpler gcode parsers take either the first or last value
11:41 AM unterhaus_: what happens when you want a diagonal line?
11:41 AM ZincBoy[CAON][m]: There are two z words on the same line
11:42 AM unterhaus_: oops, missed that
11:42 AM ZincBoy[CAON][m]: I wasn't clear in my response either 🙂
11:42 AM unterhaus_: should be an error
11:44 AM Tom_L: neither look valid to me
11:44 AM jpa-: first one is a valid command with everything after ; being a comment
11:44 AM Tom_L: the Z10 Z0 would throw an error
11:45 AM Tom_L: jpa-, maybe but it's certainly poor practice
11:45 AM Tom_L: i typically put comments in () on a line by themselves
11:47 AM skunkworks[m]: I think.. fanuc uses ; as a end if bock.. ie line feed
11:48 AM skunkworks[m]: been a while.
11:48 AM jpa-: yeah, ; is not standard, it varies
11:48 AM Tom_L: best not to use it
11:48 AM ZincBoy[CAON][m]: That is the nice thing about the gcode standard, there are so many to chose from 🙂
11:49 AM Tom_L: more a control thing than a gcode standard maybe
11:49 AM Tom_L: they want to be 'special'
11:50 AM jpa-: the RS274 standard just says "a line may have only one word beginning with that letter" - interestingly for # parameters it says "If a parameter setting of the same parameter is repeated on a line, “#3=15 #3=6”, for example, only the last setting will take effect."
11:50 AM Tom_L: and they all have some unique G & M codes
11:52 AM jpa-: funnily G0 X1 0Z1 0 is valid and means same as G0 X10 Z10
11:53 AM Tom_L: you could really fsck with somebody doing that if you felt like it
11:54 AM ZincBoy[CAON][m]: Yup. And why the recommendation is to always have a . even if there is no fractional value
11:55 AM genericum: so it only works with a newline in between ?
11:55 AM Tom_L: i put trailing zeros on
11:56 AM genericum: it has to work on a fanuc and afaik all the lines end with ;
11:56 AM unterhaus_: is the lcnc gcode parser that difficult to repurpose? I looked at the marlin g code parser and died
11:57 AM unterhaus_: I got better
11:57 AM ZincBoy[CAON][m]: You would be better off asking on a fanuc focused forum. The linuxcnc parser has enough differences that you won't get accurate responses here.
11:57 AM Tom_L: lcnc: an end of line marker (carriage return or line feed or both).
11:58 AM skunkworks[m]: linuxcnc gcode is based on a standard from nist.
11:58 AM genericum: ZincBoy[CAON][m], dont know any :(
11:58 AM ZincBoy[CAON][m]: CNCzone
11:58 AM genericum: skunkworks[m], nist sounds good
11:59 AM skunkworks[m]: https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=823374
11:59 AM skunkworks[m]: not exactly though of course..
11:59 AM genericum: skunkworks[m], ty
11:59 AM jpa-: unterhaus_: linuxcnc parser has pretty nice python bindings at least, e.g. axis uses it for the preview
12:00 PM ZincBoy[CAON][m]: If you are targeting a specific control (and you pretty much have to) then you need to look at the docs for that specific control. AFAIK nothing implements the NIST standard as written. They all have "improvements" or extensions.
12:01 PM Tom_L: https://www.nist.gov/publications/nist-rs274ngc-interpreter-version-3
12:02 PM unterhaus_: someone kept posting on the ML that the lcnc parser should be more like the fanuc parser and I suggested he was just the man to implement that
12:02 PM unterhaus_: it's easily replaced, as I understand it
12:03 PM genericum: unterhaus_, python ... I'll pass
12:03 PM ZincBoy[CAON][m]: * for certain values of easy
12:03 PM unterhaus_: the parser isn't python, it's fortran or something :)
12:03 PM genericum: It is an error if:
12:03 PM genericum: all axis words are omitted.
12:03 PM genericum: :S
12:04 PM unterhaus_: must be c
12:04 PM unterhaus_: that has to be a nist file
12:04 PM Tom_L: genericum, what is the target control for the parser?
12:06 PM unterhaus_: if you can make the parser work with fanuc, it's easy enough. Otherwise the task will be assigned to all the developers currently working on the lcnc parser. I.e., nobody
12:07 PM genericum: Tom_L, its a fanuc 18i-tb
12:08 PM Tom_L: https://www.fadalcnc.com/media/pdf/Old_Service_Files/Fanuc%20Info/18i%20-%2016i/B-63530EN_03.pdf
12:10 PM genericum: Tom_L, looks good ty
12:10 PM Tom_L: may not cover the interpreter
12:12 PM Tom_L: https://cncmanual.com/download/1505/
12:22 PM genericum: i just read another one ... says G00 is done without moving axis simultaneously ?
12:23 PM genericum: Tom_L think it was just the one :)
12:28 PM jpa-: G0 may move in something else than a straight line, theoretically "whatever is fastest"
12:28 PM jpa-: for cartesian machines straight line is usually the fastest, but for SCARA or other geometries it may be curved path
12:34 PM Rab: I crashed a Yaskawa robot control by programming straight-line rapids at max speed. Straight lines were a combination of rotating joints, and the path exceeded max velocity for individual joints.
12:38 PM genericum: the manual said simultaneous motion for 2 axis a few pages before
12:38 PM Rab: I was surprised that it failed in the way that it did (unrecoverable error requiring a hard reboot), but it was fairly elderly...maybe 1990s vintage.
12:44 PM JT-Shop: 48" box and pan brake has a new zip code
01:03 PM daight420[m]: Anybody got probe recommendations? Preffer wireless
01:05 PM jpa-: what is your budget?
01:07 PM CaptHindsight[m]: I spoke with a supposed product development "whiz" at a maker space using Centroid to retrofit a CNC lathe and mill with faulty controls ... has said that Centroid is easier and has more features than LCNC. I didn't hear more...
01:11 PM CaptHindsight[m]: https://www.centroidcnc.com/centroid_diy/acornsix_cnc_controller.html
01:12 PM TurBoss: good marketing 🤣
01:12 PM CaptHindsight[m]: looks like they use Effnix FPGA's
01:14 PM CaptHindsight[m]: TurBoss: yeah, it happens a lot to well trained consumers
01:15 PM jpa-: well, it's not strictly false either: many linuxcnc features require a lot of tinkering :)
01:15 PM jpa-: even something as basic as probing yields quite a bunch of questions
01:16 PM CaptHindsight[m]: semantics
01:17 PM CaptHindsight[m]: but this a hardware product developer, or is he?
01:18 PM daight420[m]: Ideally under 1k
01:18 PM CaptHindsight[m]: outside of the couple of trolls on the LCNC user mail list I never hear actual discussions about LCNC vs Centroid and similar
01:19 PM daight420[m]: I have a reinashaw probe but the interference is like 1800
01:19 PM daight420[m]: And I have a presenter but both need different interference units lol
01:22 PM CaptHindsight[m]: Lcvette: don't you carry a half decent probe on your website?
01:24 PM CaptHindsight[m]: daight420: good or low cost, pick one :) or look for used
01:28 PM daight420[m]: Low cost is better than none imo
01:29 PM CaptHindsight[m]: https://smallshopconcepts.com/products
01:29 PM rmu: daight420[m]: i got a vers.by wireless versaprobe and am quite happy with it. but it seems the company in belarussia or at least the webshop is gone
01:30 PM rmu: price was around €100
01:31 PM rmu: something similar to this https://www.forum.linuxcnc.org/show-your-stuff/34654-probe-from-vers-by
01:33 PM -!- #linuxcnc mode set to +v by ChanServ
01:35 PM CaptHindsight[m]: rmu: aren't they at war or something?
01:38 PM rmu: belarussia != russia, but situation there is also very "unpleasent"
01:38 PM rmu: i bought the probes around ~2018
01:41 PM daight420[m]: Jeez that's a lot cheaper than I was expecting
01:46 PM -!- #linuxcnc mode set to +v by ChanServ
01:50 PM rmu: video i just found about it. https://www.youtube.com/watch?v=zsUklQTTEz0 no affiliation either with youtube channel nor versa
01:56 PM CaptHindsight[m]: MinistryofBrokenEndMills 🤣🤣
01:58 PM CaptHindsight[m]: and tell me where you are from without mentioning the name
02:09 PM unterhaus_: doesn't clough42 have an Acorn? Whatever it was, he had to debug it
02:11 PM unterhaus_: yeah, he did and the opening shot of the video shows him hooking up a logic analyzer to the centroid
02:12 PM unterhaus_: "easy"
02:14 PM CaptHindsight[m]: ah well
02:14 PM unterhaus_: they used to be here in town, but they moved up the way an hour
02:14 PM unterhaus_: constantly advertising for programmers
02:14 PM CaptHindsight[m]: almost know it alls
02:15 PM unterhaus_: I dont' know if they are still this way, but they didn't used to answer any emails unless you gave your serial number
02:15 PM unterhaus_: not even to tell you they wouldn't answer your email without the serial number
02:16 PM unterhaus_: It took me a couple of years to find out why they didn't answer my emails, I don't think it was on their website
02:18 PM unterhaus_: I changed some settings on the centroid control at work because the programmer couldn't get the spindle to go the right speed. It wasn't bad.
02:19 PM unterhaus_: It's nice if you only deal with one kind of hardware that you buy off of ebay and sell for 4x the price
02:23 PM unterhaus_: they have that 4th/5th axis thing that I don't see on their website. Used to be pretty popular I though
02:41 PM elia: new machine day (not CNC tho): https://0x0.st/HipN.jpg
02:41 PM elia: https://0x0.st/HiJQ.jpg
02:41 PM elia: https://0x0.st/HiJq.jpg
02:42 PM elia: had to get it upright, 3t chain hoist and sling did the job
02:49 PM Rab: elia, looks like a pretty serious Polish Deckel clone?
03:08 PM -!- #linuxcnc mode set to +v by ChanServ
03:15 PM elia: Rab: yep, quite a bit bigger, 1800kg of iron
03:16 PM elia: ball screws, box ways
03:21 PM c101horse8332[m]: good evening everyone
03:22 PM -!- #linuxcnc mode set to +v by ChanServ
03:32 PM c101horse8332[m]: are lovejoy couplings competent enough for coupling a servo to leadscrew?
03:35 PM Tom_L: should be
03:36 PM Tom_L: i'm using smaller ones on mine
03:37 PM ZincBoy[CAON][m]: If properly rated and using appropriate spiders.
03:38 PM Tom_L: unless you can get perfect alignment then use a hard coupler
03:39 PM Tom_L: https://buildyourcnc.com/item/coupling-rigid-one-piece-!25-!5
03:39 PM Tom_L: those are a definite no
03:40 PM ZincBoy[CAON][m]: Those are fine if used appropriately* (* they are never used appropriately).
03:41 PM Tom_L: if you ever got a hard stall they'll snap
03:41 PM Tom_L: but yeah
03:41 PM ZincBoy[CAON][m]: The torque rating on most of those spiral ones is too low for most cnc uses.
03:41 PM Tom_L: toys
03:42 PM ZincBoy[CAON][m]: I have used them in microscope positioning units but the torque available is very small and dynamics were not really a concern.
03:44 PM ZincBoy[CAON][m]: My mill has hard couplers between the ballscrews and the servos. It has lasted 40 years so they must have got the alignment pretty good.
03:44 PM Tom_L: my sherline had them as well and i never had trouble with it
03:45 PM CaptHindsight[m]: trade solid couplings for loose bearings
03:45 PM Tom_L: it can be done just takes some patience
03:49 PM CaptHindsight[m]: c101horse🇳🇱: https://www.youtube.com/watch?v=SPnTA3H2G7g you might enjoy
03:49 PM CaptHindsight[m]: also https://www.machinedesign.com/archive/article/21828607/couplings-and-shafts
03:50 PM CaptHindsight[m]: https://fractory.com/types-of-couplings/
03:51 PM c101horse8332[m]: The one on the picture looks funky
03:51 PM Tom_L: at least it would shear
03:52 PM CaptHindsight[m]: the answer is always simple until you understand the problem
03:52 PM JT-Shop: now to give the 3 in 1 a new zip code...
03:55 PM CaptHindsight[m]: https://www.youtube.com/watch?v=H-asC_AxEyI
03:57 PM CaptHindsight[m]: https://www.youtube.com/watch?v=UuMw-i8bb-4
03:58 PM CaptHindsight[m]: https://www.youtube.com/watch?v=JILP_4GJJH0
03:59 PM Tom_L: having a YT kindof day?
03:59 PM Tom_L: CV joint
04:01 PM CaptHindsight[m]: Ass umptions
04:02 PM CaptHindsight[m]: Tom_L: rough past few decades?
04:07 PM Tom_L: not at all
04:22 PM cabbagedreams: how would i go about setting up HAL to do PID spindle control? it's probably a little strange, but i have control of a triac and zero cross input over the parallel port
04:24 PM cabbagedreams: it's a triac -> bridge rectifier -> brushed dc motor. phase control isn't optimal, but here we are
04:26 PM pcw-home: Do you have a spindle encoder
04:27 PM pcw-home: ?
04:27 PM cabbagedreams: yes
04:27 PM cabbagedreams: a and z inputs are working
04:28 PM pcw-home: Are you asking how to setup the PID?
04:29 PM cabbagedreams: i expect it'll have to be some custom HAL setup, because normally you'd have a speed controller
04:31 PM pcw-home: The PID setup would be standard for a spindle (FB from encoder RPM velocity pin, command from commanded spindle speed)
04:31 PM pcw-home: FF0 == 1
04:32 PM pcw-home: The Triac control would probably use the oneshot component
04:33 PM pcw-home: (and might need a somewhat higher servo thread rate for decent resolution)
04:35 PM cabbagedreams: reading about oneshot
04:35 PM pcw-home: a 1 KHz servo thread would only give you 8 steps, too bad there's not a base thread driven PWM comp
04:38 PM cabbagedreams: so could this be setup to trigger(it'll trigger at 120hz) when a zero cross event occurs, and then based on whether it needs more or less RMS power, it can either start conducting immediately, or wait until a later part of the cycle?
04:38 PM cabbagedreams: the need for syncronization to the power is important
04:38 PM pcw-home: Yes you set the PWM width to vary the trigger time from zero crossing
04:39 PM cabbagedreams: nice. this seems more doable than i originally thought
04:39 PM * JT[m] uploaded an image: (245KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/cXrluGIWRxdLIMHYIIjNyICI/20230311_163805.jpg >
04:39 PM pcw-home: but you have servo thread time jitter so you might want a 2 Khz or 4 Khz thread
04:40 PM JT-Shop: I'm happy with the test cut on the circle jig
04:40 PM pcw-home: That looks great
04:40 PM * JT-Shop still can't figure out servo thread percent...
04:41 PM JT-Shop: thanks
04:41 PM JT-Shop: cutting by hand was pretty bad
04:41 PM pcw-home: Is that routed?
04:42 PM JT-Shop: I'll show you in a sec
04:43 PM * JT[m] uploaded an image: (275KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/wEcSUAditnTYPJXrijOMMAFf/20230311_164225.jpg >
04:43 PM * JT[m] uploaded an image: (257KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/EIhxQjHNMANhmShXTtyXMNEs/20230311_164230.jpg >
04:45 PM boostedcabbage[m: ah nice. discord is connected with irc
04:45 PM pcw-home: JT: The calculation was correct for read/write but the servo thread calc left out the CPU speed
04:47 PM JT-Shop: I can't seem to get the right conversion from 2333 MHz to Hz or something else
04:47 PM JT-Shop: maybe the servo period of 1ms to second...
04:48 PM rmu: pcw-home: what do you mean with base thread driven PWM comp?
04:49 PM pcw-home: rmu: oneshot not pwm
04:50 PM pcw-home: sorry mixed up the terminology
04:50 PM XXCoder: boosted yep
04:50 PM XXCoder: man blondihack video so fun today
04:51 PM pcw-home: JT-Shop the servo thread time was off by CPU speed/1 GHz (so in the example, 2.3333)
04:52 PM rmu: hmm. if it's only floating point that prevents oneshot from running in base thread, that is only a limitation in kernel-mode stuff AFAIK, so in preempt-rt it should work even in base thread.
04:52 PM pcw-home: I wondered about that
04:52 PM pcw-home: might be interesting to try
04:53 PM rmu: kernel code usually couldnuse FPU
04:53 PM rmu: couldn't use FPU
04:53 PM pcw-home: Does that mean all the manual pages need updating?
04:53 PM cabbagedreams: one concern i have is the zc is a pretty short pulse
04:53 PM rmu: but if that remains true today where all kinds of vectorized stuff is uses... maybe that changed
04:54 PM * boostedcabbage[m uploaded an image: (1038KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/pUgwWraRzXtCweggBLuZWObD/PXL_20230311_220614311.jpg >
04:55 PM pcw-home: cabbagedreams: Ideally, running the Oneshot at the base thread rate should eliminate most issues
04:55 PM cabbagedreams: is that wide enough?
04:55 PM pcw-home: both in trigger detection and width resolution
04:56 PM pcw-home: how wide?
04:57 PM cabbagedreams: not very. see image above. div is 5ms
04:57 PM pcw-home: you would need > 1 ms for a 1 ms servo thread, but if you can run the oneshot in the base thread even a much narrower pulse would be fine
04:59 PM pcw-home: (note that you have to do some math to the oneshot width since longer means less output)
04:59 PM pcw-home: so the offset component at least
05:00 PM cabbagedreams: i don't know what my pp considers logic high, but at like 2v it's less than a 1ms for sure
05:00 PM rmu: https://stackoverflow.com/questions/13886338/use-of-floating-point-in-the-linux-kernel
05:00 PM pcw-home: so PID --> offset --> oneshot --> Triac
05:00 PM cabbagedreams: i guess i should just try to get the trigger setup and see
05:01 PM pcw-home: Worst case you may have to run a 4 KHz or so servo thread
05:01 PM pcw-home: Are you running Preempt-RT or RTAI?
05:02 PM cabbagedreams: 6.0.5.14.realtime1-4-rt #1 SMP PREEMPT_RT
05:03 PM pcw-home: ok so possibly you can run the oneshot/offset in the base thread, but this is not certain
05:06 PM rmu: easy enough to try. as the base thread on preempt rt is a normal linux thread, just with special scheduler hints attached, using FPU shouldn't be a problem.
05:11 PM pcw-home: halcmd: loadrt offset
05:12 PM pcw-home: halcmd: addf offset.0.update-output base-thread
05:12 PM pcw-home: HAL: ERROR: function 'offset.0.update-output' needs FP
05:12 PM pcw-home: sigh...
05:12 PM rmu: hmm
05:13 PM pcw-home: may just be out of date code in offset
05:13 PM rmu: according to that manpage https://linuxcnc.org/docs/html/man/man9/threads.9.html it should be possible to enable FP even in RTAI case
05:16 PM pcw-home: I wonder if those flags are passed in the normal hal file parsing (which may implicitly use the threads comp)
05:16 PM rmu: http://linuxcnc.org/docs/html/man/man9/motion.9.html loadrt motmod [base_period_nsec=period] [base_thread_fp=0 or 1]
05:17 PM rmu: "...base_thread_fp can be used to enable floating point in the base thread (for example for brushless DC motor control)...."
05:18 PM cabbagedreams: what is a hal net?
05:19 PM cabbagedreams: i see things like: net spindle-index encoder.0.phase-Z
05:20 PM rmu: cabbagedreams: best RTFM here https://linuxcnc.org/docs/html/hal/basic-hal.html#_net
05:20 PM pcw-home: Thanks RMU, it works!
05:20 PM cabbagedreams: rmu: indeed. thanks
05:21 PM rmu: nice. thanks. yw. i didn't do anything ;)
05:23 PM cabbagedreams: i'm just happy with how versatile the HAL is :)
05:24 PM pcw-home: No complaints and offset works (so oneshot should as well)
05:35 PM cabbagedreams: perfect. i appreciate the guidance
05:47 PM cabbagedreams: the order of creating net(s) and direction is sort of confusing me. right now i have "digital in 0" connected to the input i want to use. i should remove that net. does the chain start here, or does it really start from pid?
05:48 PM cabbagedreams: (start with oneshot instead of pid)
05:48 PM cabbagedreams: i'm probably thinking about this the wrong way
05:52 PM pcw-home: the is no order to nets except that the first token is always the signal
05:52 PM pcw-home: there
05:53 PM pcw-home: (and signals have arbitrary names)
05:53 PM pcw-home: You can think of signals as wires
05:54 PM pcw-home: net blue_wire some_source_pin some_destination_pin some_other_destination_pin
05:55 PM pcw-home: net blue_wire yet_another_destination pin
05:55 PM cabbagedreams: so it just connects the pin to a label for use?
05:56 PM pcw-home: basically connects a pin to a netname
05:57 PM pcw-home: and net order in the hal file is unimportant
05:57 PM pcw-home: (addf order _is_ important)
06:08 PM pcw-home: You might need to daisy chain two oneshot comps if you need a fixed pulse width for firing the Triac
06:21 PM cabbagedreams: i should probably turn off dynamic frequency scaling eh?
06:21 PM pcw-home: that usually improves latency
06:24 PM cabbagedreams: seem to have jitter. getting 107k on the servo thread and 56k on the base thread watching some youtube vids/browsing
06:24 PM cabbagedreams: well. max jitter
06:50 PM perry_j1987: hows it going
06:57 PM boostedcabbage[m: hello
07:11 PM boostedcabbage[m: that zc input is showing with the hal scope. this makes me happy
07:55 PM roycroft: i almost forgot - we go on summer time tonight
07:56 PM XXCoder: yep
07:58 PM roycroft: i need to go to pdx tomorrow
07:59 PM roycroft: i was going to leave early, but since i'll be losing an hour of sleep, perhaps i won't be going as early as i planned
07:59 PM XXCoder: airport you usually is better going in early than rigjt on time
08:00 PM roycroft: i'm not going to the airport
08:00 PM roycroft: pdx is a generic term for portland here in oregon
08:01 PM roycroft: the portland metro area, actually
08:01 PM roycroft: i'm really going to beaverton
08:02 PM XXCoder: ahh ok lol
08:03 PM Rowow5457[m]: Is there any documentation on setting up the voltage output for my mesa 7I96S that controls the stepper driver steps and direction voltage. It goes from -5v to +5v when I got example change direction. On my breakout board the same driver and motor work fine but it does 0-5v
08:03 PM Rowow5457[m]: It's not working at all on the mesa board
08:03 PM Rowow5457[m]: s/got/for/
08:08 PM * skunkworks[m] uploaded an image: (564KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/nZrLZVvRpKJIXOJlKJiNpKUR/PXL_20230312_020552296.MP.jpg >
08:08 PM * skunkworks[m] uploaded an image: (583KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/GUoggjSQfKqAGLOjsjCZryPi/PXL_20230312_020545769.jpg >
08:08 PM boostedcabbage[m: nice
08:10 PM boostedcabbage[m: qtdragon thinks my lathe is a mill. is there something obvious i'm missing? it's a lathe config
08:10 PM skunkworks[m]: So I should be able to make this...
08:10 PM * skunkworks[m] uploaded an image: (668KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/wIMsWSLaTWOcQYnsddfRMOjp/PXL_20230312_020905013.jpg >
08:13 PM boostedcabbage[m: should be a way to select like TTL mode
08:13 PM boostedcabbage[m: http://www.mesanet.com/pdf/parallel/7i96sman.pdf
08:16 PM * skunkworks[m] uploaded a video: (818KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/jVRDfQOrwJcPUcsHXQZPyhtX/334672460_6011740282252287_4898833235698839400_n.mp4 >
08:19 PM boostedcabbage[m: Rowow (@_discord_826552473711411210:jauriarts.org) "Step motor drives with single ended inputs... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/125271038bfaad5b2d7fe3f7a6ccb7471783480b>)
08:26 PM boostedcabbage[m: so like step+ into your driver's opto(step+), and then step -(on driver) to 7I96Ss GND
08:27 PM roycroft: url shorteners would make this channel more readable during periods of high volume link posting
08:28 PM Rowow5457[m]: What would the enable pin be
08:28 PM skunkworks: wow - that is pretty fucked up when posting on discord.. Sorry roycroft.
08:32 PM boostedcabbage[m: if the enable expects 0v or 5v, same story
08:34 PM Rowow5457[m]: So would it always be left at on?
08:34 PM Rowow5457[m]: Also thank you so much earlier for checking out the ttl mode on the manual. Idk why but this time it registered that I need to connect the - pin to ground and the + pin to the step+ pin. I instead was connecting the ground pin for step to the step- pin
08:35 PM Rowow5457[m]: I spent like 4 hours on this haha you helped something click on my brain I'm way too hungry and tired lol
08:38 PM roycroft: it's not usually a problem, but when there are 4-5 posts right after each other it gets hard to follow
08:40 PM boostedcabbage[m: you have to decide whether you want to be able to disable the driver that way. i don't know whether your driver will just be enabled if you connect nothing
08:42 PM boostedcabbage[m: i leave them disconnected on my drivers as they apparently are pulled to be enabled
08:44 PM Tom_L: skunkworks, cool
08:47 PM skunkworks[m]: Tom_L: the fit is suprisingly amazing..
08:49 PM skunkworks[m]: math is cool
08:49 PM skunkworks[m]: (and linuxcnc)
08:50 PM boostedcabbage[m: i'm getting excited about this HAL stuff
08:51 PM skunkworks[m]: cabbagedreams the sky is the limit..
08:51 PM boostedcabbage[m: i really wish i could remove this Z axis from qtdragon
08:51 PM boostedcabbage[m: er Y
08:51 PM boostedcabbage[m: s/Z/Y/
08:51 PM skunkworks[m]: smop?
08:51 PM boostedcabbage[m: oh sorry. edited
08:52 PM boostedcabbage[m: i just don't understand why it doesn't see it's configured as a lathe
08:52 PM skunkworks[m]: I don't think qtdragon is really a finished gui..
08:53 PM boostedcabbage[m: which is the most mature?
08:53 PM skunkworks[m]: lol - axis?
08:53 PM boostedcabbage[m: well yeah
08:53 PM skunkworks[m]: (I use axis)
08:54 PM skunkworks[m]: gmocoppi?
08:54 PM skunkworks[m]: spelling
08:54 PM skunkworks[m]: I think qtdragon was a proof of concept.
08:57 PM Tom_L: skunkworks[m], you sync the spindle & x or such to get that?
08:57 PM Tom_L: i forget what the comp was you did for it
08:57 PM Tom_L: custom comp
08:58 PM skunkworks[m]: Tom_L: yes - kinda spindle synced motion
08:58 PM skunkworks[m]: Wrote a comp..
09:00 PM boostedcabbage[m: i might as well go back to axis for now. explore later. need to get the lathe up
09:02 PM Tom_L: skunkworks[m], you have to write a different comp for each shape?
09:08 PM infornography: test
09:08 PM infornography: Whats with the [m]s? I don't have an [m]
09:11 PM infornography is now known as infor[n]ography
09:12 PM boostedcabbage[m: discord is connected
09:14 PM infor[n]ography: They are the discordians then
09:14 PM XXCoder: sorta
09:14 PM XXCoder: [m] is matrix
09:14 PM XXCoder: discord bridge though matrix
09:16 PM boostedcabbage[m: i feel like i'm cheating on irssi right now
09:21 PM skunkworks[m]: Tom_L: it started as a way to make polygons.. And D shapes.. Now you can 'scan' any shape. (Within tool geometry)
09:24 PM skunkworks[m]: (Single comp)
09:26 PM Tom_L: nice
09:26 PM skunkworks[m]: Cabbagedreams: I use both..
09:33 PM unterhausen: strangely, I can't connect to discord rn
09:35 PM unterhausen: does all of discord have a problem with apostrophes?
09:35 PM boostedcabbage[m: no
09:36 PM boostedcabbage[m: whatever's converting it isn't doing character set things properly
09:36 PM unterhausen: sure blame the matrix
09:37 PM boostedcabbage[m: i'm a burned out web developer. i earned it! 😄
09:41 PM roycroft: the crux of the biscuit is the apostrophe
09:43 PM boostedcabbage[m: ascii problems in a unicode world
09:49 PM unterhausen: does LCNC let you change the scale to bananas?
09:50 PM Tom_L: if you monkey with it enough you can
09:51 PM roycroft: just remember that bananas are radioactive
09:51 PM roycroft: so be careful
09:52 PM XXCoder: so are grinate desktops
09:52 PM XXCoder: countertops
09:52 PM XXCoder: fun isnt it
09:52 PM boostedcabbage[m: HAL: ERROR: function 'oneshot.0' needs FP - does this mean i can't run in the base thread?
09:52 PM XXCoder: https://xkcd.com/radiation/ always theres an applicable xkcd.
09:54 PM Lcvette[m]: cabbagedreams: have you tried probe basic lathe?
09:56 PM boostedcabbage[m: i haven't. looks cool
09:57 PM Lcvette[m]: If you are wanting a dedicated lathe interface its a potential candidate
09:58 PM xzyzz: don't let him pressure you, he wants everyong to try it :)
09:58 PM boostedcabbage[m: i'll definitely check it out after i sort out this HAL stuff
09:58 PM boostedcabbage[m: i mean, it looks pretty awesome at a glance
10:00 PM Lcvette[m]: its just another option
10:00 PM Lcvette[m]: lots of flavors is a good thing
10:00 PM xzyzz: lol just teasing you
10:01 PM boostedcabbage[m: so it looks like i can enable FP in teh base thread
10:05 PM -!- #linuxcnc mode set to +v by cadmium.libera.chat
10:05 PM -!- #linuxcnc mode set to +v by cadmium.libera.chat
10:07 PM XXCoder: ugh sounds like split to update stuff didnt go well
10:13 PM -!- #linuxcnc mode set to +vv by cadmium.libera.chat
10:18 PM Tom_L: what a pita
10:18 PM XXCoder: yeah, I suspect it was supposely to go smoothly switching out servers
10:19 PM XXCoder: but it didnt
10:19 PM XXCoder: I look "forward" to bridge reconnecting
10:19 PM XXCoder: ill ask turboss to restart server in couple hours or so, while its quietest here
10:19 PM Tom_L: worse than the netsplits on irc
10:21 PM XXCoder: some is rejoining, but thats matrix
10:22 PM Tom_L: at least i'm not logging them
10:23 PM XXCoder: lol https://www.reddit.com/r/Machinists/comments/11p4rc8/found_an_insert_at_work/
10:23 PM Tom_L: lathe
10:24 PM Tom_L: or just trickery
10:24 PM XXCoder: you'll have to pay close attention
10:26 PM Tom_L: my kid has been testing cutters at work
10:26 PM Tom_L: .070 max width of cut but full depth in Ti
10:26 PM Tom_L: insert cutters
10:27 PM XXCoder: I cant imange size of machine if that insert was its apparent size
10:27 PM Tom_L: iirc something like 40 ipm
10:27 PM XXCoder: interesting
10:27 PM Tom_L: in Ti..
10:28 PM Tom_L: testing to try and make the programs a bit more efficient
10:29 PM XXCoder: makes sense, if good parts at say 10% less time it increases parts made
10:30 PM Tom_L: they're used to taking wide shallow cuts and it's not easy for the ole timers to accept but he's convincing them
10:31 PM Tom_L: programmers are starting to trust him
10:32 PM XXCoder: if hes proving things then its good
10:32 PM Tom_L: harder when you're young
10:33 PM XXCoder: didnt do much of that myself
10:33 PM XXCoder: barely more than operator
10:33 PM XXCoder: setups and sometimes editing programs but yeah
10:33 PM Tom_L: learning catia has helped him alot
10:34 PM Tom_L: at first they wouldn't believe him... one day he told the programmer to get up and let him fix it himself
10:35 PM Tom_L: so he did and the problem cutter they'd been breaking is fixed
10:35 PM Tom_L: and it cut the time down
10:35 PM XXCoder: I bet that saved company a bundle
10:35 PM Tom_L: full depth at something like .60 width of cut on a large insert cutter
10:37 PM XXCoder: .6 to .04 width? or am I misreading Tom_L
10:37 PM Tom_L: .6 full depth
10:37 PM Tom_L: it's a large cutter
10:37 PM XXCoder: ok
10:37 PM Tom_L: .6 width
10:38 PM Tom_L: i don't remember if that part was Ti or alum
10:38 PM Tom_L: they make some big parts
10:38 PM XXCoder: ok. well while my brain is refusing to picture it well, its good that he figured it out
10:40 PM Tom_L: ok i'm out. gonna try and keep that hour's sleep
10:41 PM XXCoder: have good night
11:08 PM roycroft: i'd say discord is the word to describe this network and this channel tonight
11:08 PM XXCoder: roycroft: yeah netsplits broke bot
11:08 PM XXCoder: still 200+ to rejoin still
11:19 PM roycroft: i enjoyed this week's blondihacks
11:19 PM roycroft: quinn usually has issues when she pressure tests a boiler
11:19 PM roycroft: she had very few this time
11:19 PM roycroft: i think she's getting better at building boilers
11:21 PM XXCoder: I guess so? its first boiler I see her make so yeah lol her commentary makes it clear it wasnt first
11:22 PM XXCoder: xxcoder8265[m]: finally, I rejoined
11:35 PM mrec: I wonder did anyone try to implement a step/dir controller in an eg. STM32F4 / 180Mhz; and what was the max step rate achieved?