#linuxcnc Logs

Jul 05 2022

12:00 AM CaptHindsight[m]: in tornado alley tonight, 5 in the past hour
12:10 AM roycroft: when i lived in chicago the tornados were usually just south of where i lived
12:11 AM roycroft: but we did get a few in my town
12:29 AM roycroft: and now the firecrackers are setting off car alarms
12:52 AM -!- #linuxcnc mode set to +v by ChanServ
01:05 AM randy__: morning
01:05 AM randy__ is now known as randy
02:05 AM roycroft: they're still shooting them off like crazy here, and it's past midnight
02:05 AM * roycroft needs his beauty rest
02:09 AM XXCoder: sucks
02:10 AM XXCoder: not all shooting is good so far :(
04:08 AM pere: can the python module hal be used to connect to a running linuxcnc instance?
04:09 AM pere: <URL: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Using_HAL_In_Python > is a bit sparse on the topic.
04:12 AM pere: I would like to do something like halcmd can do, but from python.
04:25 AM robert1: pere: just try it. if you are running from a rip-environment, be sure to start your python in a shell that had rip-env set up
04:27 AM pere: I tried it, and even though my linuxcnc process is running, I can not talk to it.
04:27 AM pere: no idea what a rip environment is.
04:38 AM pere: I am trying this: import hal; hal.get_info_pins('x-output')
04:41 AM JT-Cave: morning
04:45 AM robert1: pere: rip = run in place
04:47 AM * JT-Cave had a bad night with the green apple trots all night
04:48 AM XXCoder: ow
04:49 AM pere: robert1: aha. perhaps I am doing it wrong. Lets say I want to get the value of the net/signal x-output and print it in python. How do I do this?
04:50 AM pere: Ah, this seem to work: import hal; hal.get_value('x-output')
04:51 AM JT-Cave: pere, feel free to add pid to the documents
04:52 AM pere: JT-Cave: I do not yet undestand it enough to give it a try.
04:53 AM JT-Cave: I don't work on the docs anymore too much has changed
04:58 AM JT-Cave: I don't understand it enough either, I just take notes when Peter speaks
04:58 AM pere: hehe. :)
04:59 AM pere: do you know why you do not recommend touching I and D in the PID controller, but instead suggest to use FF1 and FF2 (and not FF0)? I have no idea what FF# is.
05:00 AM JT-Cave: like I said I just take notes when Peter speaks to help others
05:02 AM robert1: pere: FF means "feed forward". it just adds the set-point times FF1 faktor to the output.
05:03 AM Tom_L: morning
05:03 AM robert1: pere: imagine there are no errors or disturbances, then the set-point and the output should be identical (given proper scaling of set-point and output)
05:04 AM pere: robert1: thanks.
05:05 AM pere: the FF# values are not mentioned in the linuxcnc pid-theory document, which is confusing.
05:07 AM JT-Cave: a lot of people worked on the docs over the years...
05:07 AM robert1: https://linuxcnc.org/docs/html/man/man9/pid.9.html
05:07 AM robert1: sorry, that should have been FF0 in my comment above
05:08 AM pere: JT-Cave: sure. and these days I am sending fixes when I believe I understand what I talk about. :)
05:08 AM robert1: FF1 is the factor for the time derivative of the command value (set point), so in a position loop for a velocity controlled servo, you usually have FF1 = 1, no FF0, that put commanded speed to output.
05:10 AM robert1: https://forum.linuxcnc.org/forum/38-general-linuxcnc-questions/31577-pid-tuning-ff1, see last comment
05:14 AM * JT-Cave feels a tiny bit better after drinking 20 oz. of water to rehydrate
05:16 AM pere: But this fail to work: python3 -m hal -c "print(hal.get_value('x-output'))"
05:16 AM pere: robert1: thank you.
05:21 AM JT-Cave: 10 more minutes and it's time to let the chickens out and go back to bed
05:22 AM XXCoder: re-night
05:22 AM JT-Cave: I hope I can get some sleep
05:23 AM JT-Cave: pere, do you have a pin or signal x-output?
05:25 AM robert1: pere: any errors?
05:26 AM pere: it is a signal. no errors, just return immediately with no output.
05:28 AM JT-Cave: pere, is this syntax in the docs somewhere?
05:28 AM pere: JT-Cave: the -m or -c syntax? man python3?
05:28 AM JT-Cave: hal.get_value
05:29 AM pere: JT-Cave: I found it running python interactively, importing hal and then typing hal.<tab>.
05:50 AM robert1: pere: I don't think -m does what you think it does.
05:51 AM pere: you are right. I confuse it with perls behaviour.
05:52 AM pere: this seem to work: python3 -c "import hal; s = hal.component('ini'); print(hal.get_value('x-output'))"
05:52 AM pere: No idea why the s= statement is needed.
05:52 AM pere: 'ini' is picked at random as a component I believe exist.
05:55 AM JT-Cave: https://pastebin.com/A6Qwj3qj
05:56 AM JT-Cave: that code will poll the linuxcnc.stat() and print out everything
05:56 AM * JT-Cave wanders up to the couch
05:59 AM pere: JT-Cave: thanks. lots of useful info there.
08:50 AM JT-Shop: pere http://linuxcnc.org/docs/devel/html/config/python-interface.html
09:35 AM pere: Perhaps <URL: https://github.com/LinuxCNC/linuxcnc/pull/1801 > make the pid component easier to understand from reading the manual page.
09:39 AM roguish[m]: pere: how does that PID man page look in real man page?
09:39 AM ZincBoy[CAON][m]: http://linuxcnc.org/docs/stable/html/motion/pid-theory.html Does a pretty good job? I though the man page for PID was pretty good in describing what it does and what the inputs/outputs are.
09:40 AM pere: very nice.
09:41 AM pere: ZincBoy[CAON][m]: I assume you have a good answer to my questions about the terms used there, the one I posted here in IRC earlier.
09:41 AM roguish[m]: are the actual calc shown anywhere? like written out, not just in criptic C
09:41 AM pere: I sure do not, so for me that page do not impress much. it is useful, sure. :)
09:42 AM ZincBoy[CAON][m]: Sorry, what terms?
09:42 AM pere: roguish[m]: I cut-n-pasted the description from the C comment. to me it is crystal clair calculation. what do you miss and how would you want it in the manual page? I for sure welcome better patches than mine.
09:43 AM pere: ZincBoy[CAON][m]: 'Rise time'? and 'determine critical gain'
09:44 AM roguish[m]: there is a block diagram somewhere. be good to link to it.
09:44 AM pere: the page do not explain these terms, and they seem to be important for the understanding.
09:45 AM ZincBoy[CAON][m]: Google is your friend. Those are generic terms and not something that should need to be defined.
09:48 AM pere: ZincBoy[CAON][m]: so it is as I suspected, you understand the text while I do not, and thus believe the text is pretty good.
09:50 AM ZincBoy[CAON][m]: Yes, I have read about control systems and the equations behind them. You won't understand how this works in detail without delving into the background. The LinuxCNC docs won't replace a textbook.
09:50 AM * roguish[m] posted a file: (862KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/HmZVYDbnqiaejQhGUQVvejkN/EMC2_Manual_Pages%20(2).pdf >
09:50 AM roguish[m]: pere: check this out. go to page 131.........and look at the date
09:52 AM pere: roguish[m]: I saw a date. now what?
09:52 AM * roguish[m] uploaded an image: (90KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/VTVBFWybaxNTPCraJiudFIpd/image.png >
09:53 AM roguish[m]: pere: just for info.. goto page 131. PID
09:54 AM pere: roguish[m]: you find that diagram clearer than the textual description of the calculation?
09:54 AM roguish[m]: I think they work together pretty well.
09:55 AM pere: sound like you should send a patch to link to it from the manual page, and make sure the diagram is listed in the pid-theory.adoc text.
09:56 AM pere: roguish[m]: My takeway from page 131 is that too few people improve the documentation of the linuxcnc project. what that your point?
09:56 AM roguish[m]: I would like to , but have no idea at all on how. if i did anything, it would for sure just mess it all up.
09:56 AM pere: roguish[m]: want to learn?
09:56 AM roguish[m]: not right now. but maybe in the future.
09:58 AM roguish[m]: you are correct. too few people work on docs. also, too few people read docs. i, personally, am a firm believer in RTFM
10:00 AM ZincBoy[CAON][m]: It is the paradox of documentation. Those that know enough to write the docs don't need them so don't see the value. Those than need the docs can't write them but see lots of value. The only way this gets fixed is with a unicorn of someone skilled who enjoys writing documents or lots of $$$.
10:01 AM ZincBoy[CAON][m]: LinuxCNC is actually pretty well documented. I have been able to get systems running purely by RTFM.
10:12 AM pere: ZincBoy[CAON][m]: do you think the current status is a sensible bar of entry, ie those with your skill level should make it, others should not?
10:27 AM ZincBoy[CAON][m]: I stand by LinuxCNC being well documented. I started out with zero CNC knowledge and was able to get machines running (back when LinuxCNC was still EMC2). There are some really good tools to create basic configs that 99% of users will find sufficient. Anyone going beyond than and playing with industrial control systems better have a good understanding of what they are doing. This goes beyond the LinuxCNC docs. Being able to
10:27 AM ZincBoy[CAON][m]: research things you don't understand out side of the docs is a foundational skill. If you don't have that, then I think that should be a bar to entry. Industrial class systems will kill you without even trying and the margin for error is low.
10:29 AM CloudEvil: And require a lot of fundamental knowledge of their own as absolute prerequisites.
11:00 AM pere: It make me sad to see such elitism here, apparently used as an argument to not improve the documentation. I hope that is not intended.
11:01 AM pere: on a more interesting note, I got the carousel component working just now. not quite sure how to hook it into the linuxcnc machinery yet, but can select tool and get it into the correct location.
11:01 AM roguish[m]: pere: I found that PID block diagram in the current docs. but it's virtually unreadable. very poor resolution... http://linuxcnc.org/docs/stable/html/hal/rtcomps.html#sec:pid
11:02 AM roguish[m]: Where is the original art? any clue?
11:05 AM roguish[m]: found it..........
11:16 AM CloudEvil: I am not arguing that the documentation should not be improved.
11:17 AM cpresser: pere: I don't think its elitism. Control-Theory is something one can study. Like for 5 years at a university. Putting all that info the LinuxCNC doc does not sound right to me. What could be done however is add links to the respective resources where one can read up on those topics.
11:17 AM pere: in this context that is how I understand it. good to hear that is not your intention.
11:17 AM cpresser: Its definitely not the intention to have shitty docs
11:18 AM pere: well, the feedback I just got for reading the doc and asking questions about the unexplained terms used was that the description is pretty good...
11:19 AM cpresser: I am also biased, since I know the terms in this context. But, I feel the same when reading legal text. I have no clue about quite a few of the terms used.
11:21 AM cpresser: Given the complexity of the topic, I think it is fair to expect people to just google the terms and start reading wikipedia or something.
11:22 AM cpresser: Adding links to those resource, into the existing docs, to make that reading easier would slightly improve the situation.
11:22 AM pere: well, I believe a text should be self contained, with explicit references to external sources if that is intended.
11:24 AM cpresser: Its not elitism, but people not bothering to put even more work into docs. I rather spend my time tinkering with $project.
11:27 AM roguish[m]: LinuxCNC is a difficult one to document. It's sophisticated enough for professional use, but at the same time is very much a DIY system. covers a very wide range of users.
11:28 AM ZincBoy[CAON][m]: It is this. Those terms are so foundational that defining them would be odd in something like the LinuxCNC docs.
11:28 AM pere: well, my experience from reading docs and sending patches to improve the text do not really match that experience.
11:32 AM ZincBoy[CAON][m]: No text can be self contained. They are all based on context. There are simply different levels of context required to understand a given text. If you fall below that threshold then you need to educate yourself to gain the context needed. That is where google or textbooks come in to help you gain that understanding.
11:35 AM Tom_L: the docs have improved immensely over the years and i've followed it since it was a nist project
11:35 AM Tom_L: nowdays a noob really has no excuse to get something working
11:36 AM pere: then I guess it is good enough and I should stop sending patches.
11:37 AM Tom_L: it has improved alot even since i started a build and that work is appreciated
11:37 AM roguish[m]: pere: how can the PID block diagram be updated to a better (higher) resolution image? there are 3 formats: png, eps, svg, and a dxf with a different name
11:37 AM Tom_L: this documentation push of late is greatly appreciated at least by me
11:38 AM roguish[m]: Tom_L: totally. it's great
11:38 AM pere: sad to report that the ddt component is not doing a very good job at calculating the derivative. :(
11:39 AM pere: roguish[m]: I suspect the DXF comments in docs/README.adoc might have some background information. Discovered that section while translating.
11:42 AM roguish[m]: reading it now. thanks
11:46 AM roguish[m]: pere: I can't tell which of the 4 file types is used in the actual docs. pdf, html, and man. do you know?
11:47 AM pere: roguish[m]: the question do not make sense to me. most derive from the same .adoc source files. some man pages are generated, most are manually written. I hope the manually written manual pages can be converted to .adoc too, and the man pages generated from adoc.
11:47 AM pere: roguish[m]: what is 'the actual docs'?
11:48 AM roguish[m]: the pdf, and html
11:49 AM pere: roguish[m]: to me that is a strange classification. I would classify in source and derived docs, and let the doc building handle all the derived files.
11:50 AM roguish[m]: ok, what is the 'source' file?
11:52 AM pere: it differ for individual document fragments. most is in docs/src/ as adoc files. some are in docs/man/ as manual pages (but mixed with generated files), and some are generated from .comp files in src/hal/components/. There is also some mostly internal .adoc files directly in docs/.
11:52 AM roguish[m]: wow. messy
11:52 AM pere: I believe all generated manual pages should have a header at the top stating that they are generated, to give people a hint to not edit them there.
11:53 AM pere: roguish[m]: I discovered it like this last authum.
11:55 AM pere: is at_pid intended as a in place replacemt for the pid component? The former lack the error-previous-target pin. :(
12:02 PM CaptHindsight[m]: how about a document that explains how to edit the documents?
12:03 PM pere: CaptHindsight[m]: docs/README.adoc and friends is not enough?
12:05 PM CaptHindsight[m]: technical documents should be written for the new user and the info should also cover the technical details that the experts require
12:07 PM CaptHindsight[m]: people that already know, well already know, they often just need a reminder or clarifications on changes
12:08 PM CaptHindsight[m]: it hard enough to do on funded commercial projects and they hire technical editors
12:08 PM CaptHindsight[m]: so open source will just end up with whatever the dev feels like at the time
12:13 PM roguish[m]: pere: I found in rtcomps.adoc the reference to pid-block-diag.png if I create a higher resolution pid-block-diag.png file, how can it get put into the source?
12:32 PM pere: roguish[m]: clone the linuxcnc git repo, create a new git branch for the change, push it to github and create a pull request for it.
12:32 PM pere: could not quite get at_pid to stop overloading the axis controller. :(
12:35 PM JT-Cave: those who think about complaining about the documents should work on the first...
12:42 PM JT-Cave: roguish[m], maybe you just need to regenerate if from the .svg or .eps file
12:44 PM JT-Cave: the .eps file is clear
12:51 PM roguish[m]: JT-Cave: I have made one from the dxf. Easiest for me
12:54 PM roguish[m]: and in no way am I bitching about the docs. I am actually trying to help, at least a little bit.
12:55 PM pere: roguish[m]: perhaps 'man at_pid' should warn users...
12:56 PM roguish[m]: maybe. it's a pretty obscure comp
12:56 PM roguish[m]: but a simple warning could help
01:01 PM pere: I tested and it mostly just shook my machine and overloaded the motor controller.
01:05 PM roguish[m]: probably not good to use.
01:06 PM roguish[m]: it's too bad. Some drives actually have very sophisticated auto-tuning, and It would be really handy.
01:44 PM atomic[m] is now known as sqishi8675[m]
02:22 PM JT-Cave: roguish[m], it was not for you
02:53 PM Tom_L: what a night. pets hate noise and kept me up a good part of the night
02:53 PM XXCoder: that sucks
02:53 PM XXCoder: here we barely had any fireworks, BIG fee if caught with em
02:54 PM XXCoder: my bros dog was scared but not too bad
02:54 PM Tom_L: nobody follows the guidelines here
02:54 PM XXCoder: just hope nobodys dumb enough to shoot up into sky
02:54 PM Tom_L: was finally able to take them out around 3
02:54 PM XXCoder: expecially auto fire
02:55 PM Tom_L: some were clearly commercial grade
03:11 PM -!- #linuxcnc mode set to +v by ChanServ
03:11 PM -!- #linuxcnc mode set to +v by ChanServ
03:37 PM -!- #linuxcnc mode set to +v by ChanServ
03:37 PM -!- #linuxcnc mode set to +v by ChanServ
03:38 PM JT-Cave: lg
03:39 PM JT-Cave: what a night, green apple trots kept me up all
03:39 PM JT-Cave: roguish[m], https://github.com/jethornton/gcode
03:47 PM _unreal_: and started the print for the last part of this cyclone
03:47 PM _unreal_: ugh 6 hours
03:52 PM tcurdt: when the enable button is disabled it means I have not all halui nets correctly connected, right?
04:00 PM _unreal_: enable disabled could mean a number of things
04:01 PM _unreal_: do you have any signals like end stops
04:03 PM tcurdt: right now it's just a very minimum config https://gist.github.com/tcurdt/e9e6519bc24590627f2c7f01f19ecaa3
04:05 PM _unreal_: END limits,ESTOP, stuff like that I would check right off the bat.
04:06 PM _unreal_: typically if you can not enable. a setting may be inverted
04:06 PM _unreal_: high/low setting input
04:06 PM tcurdt: right now nothing is connected ... I talking just about the button in the ui
04:06 PM _unreal_: is it setup with RT
04:07 PM tcurdt: you mean the kernel? it's based on the standard ISO for 2.8.2
04:08 PM JT-Cave: tcurdt, # Standard I/O Block - EStop, Etc
04:08 PM JT-Cave: # create a signal for the estop loopback
04:08 PM JT-Cave: net estop-loopback iocontrol.0.emc-enable-in <= iocontrol.0.user-enable-out
04:08 PM _unreal_: JT-Cave, you'll likely have a better time solving it then I will. been to long since I've had lcnc up and running. and its sitting right behind me :/
04:09 PM _unreal_: still waiting on the new cnc to be up and running.
04:09 PM _unreal_: ugh... god this print is going to take for ever
04:09 PM _unreal_: OMG. its going to finish at midnight maybe 1am
04:09 PM tcurdt: hm .. I just loaded a sample config and it also has the button disabled :/
04:10 PM tcurdt: sim.axis / aixs
04:10 PM _unreal_: brb time for a beer
04:11 PM _unreal_: corona :) extra. dont do that DIET SHIT.
04:12 PM JT-Cave: tcurdt, which button in axis sim?
04:12 PM tcurdt: JT-Cave "toggle machine power"
04:13 PM JT-Cave: are you out of estop?
04:13 PM JT-Cave: you must toggle estop first
04:13 PM tcurdt: aaaaaah ... ok
04:13 PM JT-Cave: :)
04:16 PM _unreal_: palm to face. I think that was one of the first things I said was check limits and estops
04:25 PM t4nk_freenode: you're a real fix-it man, _unreal_ :b
04:34 PM tcurdt: _unreal_ "check limits and estops" sounds like something very different (to me)
04:34 PM t4nk_freenode: hehe
04:34 PM t4nk_freenode: no kidding :)
04:57 PM -!- #linuxcnc mode set to +v by ChanServ
05:52 PM tcurdt: on a gantry machine with a two motors on the y, both y axis should share the same joint, right?
05:55 PM Tom_L: A gantry machine with one motor on each of two of the axes, and two motors on the third axis, has 4 joints.
05:56 PM Tom_L: http://linuxcnc.org/docs/stable/html/config/ini-config.html#_joint__lt_num_gt_section
06:02 PM _unreal_: tcurdt, dual motor on the Y then should have two limits
06:02 PM tcurdt: _unreal_ sure
06:03 PM _unreal_: single motor tied with a timing belt to dual leads on the Y should have a single limit
06:03 PM Tom_L: https://forum.linuxcnc.org/49-basic-configuration/33079-how-to-2-or-more-motors-on-one-axis-gantry-linuxcnc-2-8-master
06:03 PM Tom_L: read that
06:03 PM tcurdt: it seems like it depends on "kinstype=b"
06:03 PM _unreal_: I prefer to have single motor with timing belt.
06:03 PM _unreal_: its FIXED hard
06:04 PM Tom_L: last link shows an example
06:04 PM _unreal_: sorry was away yacking to/with my father about theme parks here in florida
06:05 PM _unreal_: I'm considering going to bush gardends. my friend wants to go
06:06 PM tcurdt: darn .. I just got the axis and joint mixed up ... so it should be 4 joints, but 3 axis
06:06 PM Tom_L: yes
06:13 PM _unreal_: in lcnc yes 4 joints 3 axis. one is mirror
06:14 PM tcurdt: with the PIDs it seems I could use names "loadrt pid names=pid.x,pid.y1,pid.y2,pid.z" but it seems "motmod" does not support naming?
06:15 PM Tom_L: there is a way but i'm not familiar how you address the 2nd motor
06:15 PM Tom_L: it may be something like Y and -Y
06:15 PM Tom_L: check the documentation
06:15 PM Tom_L: that's why it was written :)
06:15 PM tcurdt: that's what I did ... I didn't see it in there
06:16 PM tcurdt: http://linuxcnc.org/docs/html/man/man9/motion.9.html
06:16 PM Tom_L: (we were discussing the value of the docs earlier)
06:16 PM Tom_L: it is there somewhere but i'm not sure where
06:16 PM _unreal_: I would just dbl the output from the mesa to the motor contollers
06:16 PM _unreal_: for dual y
06:16 PM _unreal_: JUST ADD high value resistors
06:16 PM Tom_L: _unreal_, have you set one up?
06:16 PM t4nk_freenode: *sigh*
06:17 PM _unreal_: no I run belt drive
06:17 PM Tom_L: linuxcnc will support this but i'm not familiar with how since i don't have a gantry
06:17 PM tcurdt: under bugs it says "This manual page is incomplete." :)
06:17 PM _unreal_: My laser etcher is dual Y but that is HARD wired output from motion controller to dual Y motor drivers
06:18 PM _unreal_: but my laser etcher is also setup with GRBL but its A HARDWARE level setup
06:18 PM Tom_L: 2.8 started supporting gantry configs
06:19 PM tcurdt: if I ever get there I would like to use auto-squaring ... I'd guess that needs a separate y config
06:19 PM Tom_L: 2 limits
06:19 PM Tom_L: and yes it will do that
06:19 PM Tom_L: once the switches are set correctly
06:20 PM _unreal_: tcurdt, thats why I'm building my current cnc machine with a timing belt setup for Y
06:20 PM _unreal_: dual Y lead screw but one motor
06:20 PM _unreal_: if its out of square I have bigger issue.
06:20 PM _unreal_: issues.
06:20 PM Tom_L: you still have to get it aligned
06:21 PM Tom_L: and wonder about belt stretch
06:21 PM _unreal_: belt drive is not an issue
06:21 PM _unreal_: if the belt looses tension its going to skip teeth. thats a bigger issue
06:21 PM t4nk_freenode: I would have to fire up my rpi, I got some configs on there I used with that
06:22 PM t4nk_freenode: but I'm about to crash into a coma
06:22 PM _unreal_: NOOOOOOO......................
06:22 PM Tom_L: well get off your butt...
06:22 PM _unreal_: t4nk_freenode, O
06:22 PM _unreal_: O
06:22 PM _unreal_: I'm 40% done printing the last part for my new dust seperator.
06:23 PM _unreal_: eerrrrrr more like 30%
06:23 PM Tom_L: tcurdt, http://linuxcnc.org/docs/stable/html/config/ini-homing.html
06:23 PM Tom_L: read about home_sequence. they mention some about it there
06:24 PM Tom_L: A negative HOME_SEQUENCE also applies to commands to home a single joint. If the HOME_SEQUENCE value is negative, all joints having the same absolute value of that HOME_SEQUENCE will be homed together with a synchronized final move. If the HOME_SEQUENCE value is zero or positive, a command to home the joint will home only the specified joint.
06:25 PM Tom_L: 3rd example on that page
06:25 PM tcurdt: do I understand correctly that some sections/values in the ini are mandatory? while others are completely user defined?
06:25 PM Tom_L: you can define sections yes
06:25 PM Tom_L: and refer to them in the hal file
06:26 PM tcurdt: Tom_L that's clear ... but it seems some are also required
06:26 PM tcurdt: which is a bit weird
06:26 PM Tom_L: yes
06:26 PM Tom_L: hal looks for them
06:28 PM Tom_L: An example simulation config (gantry_jjog.ini) that demonstrates joint jogging when using negative home sequences is located in the: configs/sim/axis/gantry/ directory.
06:28 PM Tom_L: if you install that sim you can see it
06:29 PM Tom_L: 2.8 has simplified setting up a dual motor axis gantry
06:32 PM tcurdt: what was changed in 2.8 in that regard? just curious
06:32 PM _unreal_: EVERYTHING....
06:32 PM _unreal_: kidding no idea
06:33 PM Tom_L: alot of things were changed in 2.8
06:33 PM Tom_L: it was a rather lengthy list
06:33 PM Tom_L: prior to that you had to figure gantry out on your own
06:33 PM tcurdt: not sure whether to look forward to the 2.9 release then or not :)
06:33 PM _unreal_: I think thats what MY current build is based on is 2.8
06:33 PM _unreal_: but again its been some time since I have even booted it
06:34 PM Tom_L: 2.9 looks nice from what i have seen so far
06:34 PM _unreal_: tcurdt, if your dealing with a new build then yes
06:34 PM _unreal_: I would IMPRINT on the newest
06:34 PM Tom_L: newest stable
06:34 PM Tom_L: unless you require master
06:34 PM _unreal_: ^^^^^^^^^^^^^^^^^^^^^^^^^
06:34 PM tcurdt: newest stable is still 2.8.2, no?
06:34 PM Tom_L: yes
06:37 PM tcurdt: not sure I am eager to build it myself ... just saw a new UI in 2.9 that looked nice
06:38 PM MikeM[m]: Good evening
06:39 PM MikeM[m]: Silly question. Is there a way to set the "SUBROUTINE_PATH" in the INI so that PNC Config doesn't over write it?
06:40 PM MikeM[m]: My searches are leading me to believe the answer is no, but wanted to check here to make sure I am not missing somthing.
06:40 PM MikeM[m]: s/somthing/something/
07:04 PM roguish[m]: _discord_854784589033898004[m]: copy n paste
07:04 PM Tom_L: MikeM[m], no
07:05 PM MikeM[m]: That is what I thought, thanks!
07:05 PM Tom_L: pncconfig overwrites and starts over
07:05 PM Tom_L: have you looked at mesact?
07:05 PM Tom_L: it at least will let you reread the config prior to manual changes
07:06 PM Tom_L: once you get a basic config going you should wean yourself off config tools anyway
07:06 PM Tom_L: that's where all the fun is at
07:07 PM MikeM[m]: Yeah, I have been doing more with the 'custom' files manually. I setup a bunch of pyvcp buttons just to learn more about it.
07:07 PM Tom_L: using axis gui?
07:07 PM MikeM[m]: Now attempting to setup the Z part of an XYZ touch plate
07:07 PM MikeM[m]: Yes
07:08 PM Tom_L: i use 2 configs and the screens look like this: http://tom-itx.no-ip.biz:81/~webpage/cnc/screenshots/Current_Linuxcnc_Axis_screen.png
07:08 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/screenshots/Rotary_Axis_screenshot.png
07:08 PM roguish[m]: PNCCONFIG is just an aid. best to use it as a starting point, and learn from it.
07:09 PM Tom_L: i learned by reading other config files
07:09 PM Tom_L: never used a config tool
07:09 PM Tom_L: but the tools work alot better now
07:10 PM MikeM[m]: Cool
07:11 PM MikeM[m]: I will have to dive into that more deeply.
07:11 PM MikeM[m]: For Cycle Times is Current the elapsed time, and Overall the ETA time for completion?
07:12 PM Tom_L: yes including manual tool changes etc
07:12 PM Tom_L: current is per tool
07:12 PM MikeM[m]: Nice
07:12 PM Tom_L: i don't use it as much as i thought i would since my cad cam estimate is very close
07:12 PM MikeM[m]: That looks like that would come in very handy!! I will have to look into that next
07:13 PM Tom_L: my configs are on that site
07:13 PM MikeM[m]: Will take a look, thank you!
07:13 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/configs/NEW_MILL_SHOP_2.8.2/
07:14 PM Tom_L: they're kindof a mess
07:14 PM Tom_L: any _A is for the A axis config
07:14 PM Tom_L: i tend to experiment more with my configs and leave unused code
07:27 PM tcurdt: I am getting a "limit switch error" with: net x_home => joint.0.home-sw-in, net x_min => joint.0.neg-lim-sw-in, net x_max => joint.0.pos-lim-sw-in
07:28 PM tcurdt: not sure I understand the error message
07:37 PM MikeM[m]: YAY, got the a button for the Z Touch plate working!!
07:42 PM CloudEvil: :)
07:45 PM MikeM[m]: Now to work on a script for doing an ZXY touch-off 🙂
07:57 PM Tom_L: tcurdt, A. did you back off the limit when homing. B. it's wired backwards
07:58 PM Tom_L: (polarity)
08:01 PM tcurdt: Tom_L so far nothing is connected ... so I would expect the homing to fail, but I am getting this when I just toggle the power in linuxcnc
08:01 PM tcurdt: can this be due to the home config values in the ini?
08:01 PM Tom_L: no switches wired.. that's normal
08:02 PM Tom_L: the switch needs to close that circuit
08:03 PM Tom_L: trying to recall.. you may be able to bypass that with NO_FORCE_HOMING or something like that temporarily
08:03 PM Tom_L: read the home section of the docs
08:03 PM Tom_L: and ini home settings
08:03 PM tcurdt: I did
08:03 PM Tom_L: per joint
08:04 PM Tom_L: i get errors if i don't have switches connected
08:04 PM Tom_L: you want an open switch condition to trip
08:05 PM Tom_L: in case a wire is broken etc
08:06 PM tcurdt: inverting the inputs made it work ... just as a temp work around
08:08 PM Getty4 is now known as Getty
08:13 PM Tom_L: figured it was one of those 2
08:26 PM tcurdt: trying qtdragon I am getting: "STATUS.connect('max-velocity-override-changed',... must be in range" ... but can't find such a config value
08:32 PM bjorkintosh: so, I love and use cimcoedit's backplot feature a lot. is there a good gnu/opensource backplotter available?
08:46 PM CaptHindsight[m]: bjorkintosh: something like LCNC simulator?
08:47 PM bjorkintosh: yeah
08:49 PM CaptHindsight[m]: A) Install LCNC, B) pick a SIM and open, C) run your program, D) enjoy the backplot
08:49 PM bjorkintosh: hahah
08:49 PM bjorkintosh: duh. yes of course.
08:49 PM XXCoder: vismach is awesome too
08:50 PM XXCoder: I made tutorial on how to make your own vismach few years ago
08:50 PM bjorkintosh: XXCoder, http://linuxcnc.org/docs/html/gui/vismach.html
08:50 PM bjorkintosh: ?
08:50 PM XXCoder: was anything added to vismach since then?
08:50 PM XXCoder: yup, one I made is 3axis_tutorial
08:53 PM XXCoder: getting close to 4 years already. https://github.com/LinuxCNC/linuxcnc/blob/master/configs/sim/axis/vismach/3axis-tutorial/3axis-tutorial.hal
08:54 PM XXCoder: tutorial is in https://github.com/LinuxCNC/linuxcnc/blob/master/configs/sim/axis/vismach/3axis-tutorial/3axis-tutorial.ini
08:55 PM XXCoder: oops wrong file AGAIN
08:55 PM XXCoder: ok triple confirmed this time LOL https://github.com/LinuxCNC/linuxcnc/blob/master/configs/sim/axis/vismach/3axis-tutorial/3axis-tutorial
09:25 PM _unreal_: fuck my left what the hell? I just woke up. dont know how but I passed out right after my last post.
09:25 PM _unreal_: I had shit to do....
09:25 PM _unreal_: left=life
09:25 PM bjorkintosh: that's unreal!
09:29 PM _unreal_: I woke up because my 3d printer vibrated something off the cabinet its sitting on. GOD DAMN IT.
09:29 PM _unreal_: this sucks.
09:29 PM XXCoder: just randomly zonking out is bit more of a concern
09:30 PM bjorkintosh: why? fatigue guarantees that.
09:30 PM bjorkintosh: I came to a start because someone was snoring rather loudly while I was watching some tutorials online.
09:31 PM _unreal_: XXCoder, try working outside in south florida.
09:31 PM bjorkintosh: i'd simply passed out and snored myself awake.
09:31 PM _unreal_: all day
09:31 PM XXCoder: no, I mean, you didnt realize you fell asleep?
09:31 PM bjorkintosh: yes. the heat wears you out.
09:34 PM _unreal_: I'm more pissed off because I had stuff to do. its almost 11pm AND I'm 99% done with the spool on my 3d printer.
09:35 PM _unreal_: I doubt I have enought to finish the print. its 99 ish percent done.
09:35 PM _unreal_: enough
09:36 PM XXCoder: have another roll? usually printer can be paused and filament swapped
09:36 PM _unreal_: yes I have on hand its just the FUCKING risk.
09:38 PM XXCoder: well still better than simply let it fail
09:39 PM _unreal_: I was supposed to start milling the last major parts for the cnc machine
09:39 PM XXCoder: if last of filament is getting near exrtuder pause and swap
09:42 PM bjorkintosh: sorry unreal. best solution I've found is to set a timer or alarm just in case I fall asleep.
09:43 PM bjorkintosh: I learnt that the hard way when I woke up a good 45 minutes after my lunch was over some time back.
10:05 PM _unreal_: fianlly got the new filiment fed in
10:07 PM _unreal_: and in my infinite wisdom I decided I HATED having the joiner tool sitting on my desk. guess what I cant find and just spend 15min looking for a lighter for.
10:07 PM bjorkintosh: did it fall off the table?
10:08 PM _unreal_: no I HATED IT sitting on my desk. tiny device. so I moved it to a "SAFE" place where I dont have to look at it. I remember vividly picking it up.
10:08 PM _unreal_: and thats all I remember of it.
10:27 PM bjorkintosh: where do you keep small things you don't wanna lose?
10:30 PM _unreal_: there glued together all fucking done. now I'm going to take a shower FINALLY and eat din so I can wake up again and go to work and have gotten nothing done.
10:31 PM bjorkintosh: welcome to the matrix
10:31 PM _unreal_: what a waste of time. at least My cyclone is now done. if I have time. or rather so long as I dont fall asleep FIRST I'll snape a photo of it and post it.
10:31 PM _unreal_: nothing like a bit of CA glue and insta cure spray